[House Hearing, 118 Congress]
[From the U.S. Government Publishing Office]


                       VETERANS HEALTH INFORMATION
                         SYSTEMS AND TECHNOLOGY.
                          ARCHITECTURE (VISTA)

=======================================================================

                                HEARING

                               BEFORE THE

                SUBCOMMITTEE ON TECHNOLOGY MODERNIZATION

                                 OF THE

                     COMMITTEE ON VETERANS' AFFAIRS

                     U.S. HOUSE OF REPRESENTATIVES

                    ONE HUNDRED EIGHTEENTH CONGRESS

                             FIRST SESSION
                               __________

                         TUESDAY, MARCH 7, 2023
                               __________

                            Serial No. 118-2
                               __________

       Printed for the use of the Committee on Veterans' Affairs
       
       
                  [GRAPHIC NOT AVAILABLE IN TIFF FORMAT]       


                    Available via http://govinfo.gov
                                        
                               __________

                    U.S. GOVERNMENT PUBLISHING OFFICE
                    
51-536                    WASHINGTON : 2023                       
                    
                    
                     COMMITTEE ON VETERANS' AFFAIRS

                     MIKE BOST, Illinois, Chairman

AUMUA AMATA COLEMAN RADEWAGEN,       MARK TAKANO, California, Ranking 
    American Samoa, Vice-Chairwoman      Member
JACK BERGMAN, Michigan               JULIA BROWNLEY, California
NANCY MACE, South Carolina           MIKE LEVIN, California
MATTHEW M. ROSENDALE, SR., Montana   CHRIS PAPPAS, New Hampshire
MARIANNETTE MILLER-MEEKS, Iowa       FRANK J. MRVAN, Indiana
GREGORY F. MURPHY, North Carolina    SHEILA CHERFILUS-MCCORMICK, 
C. SCOTT FRANKLIN, Florida               Florida
DERRICK VAN ORDEN, Wisconsin         CHRISTOPHER R. DELUZIO, 
MORGAN LUTTRELL, Texas                   Pennsylvania
JUAN CISCOMANI, Arizona              MORGAN MCGARVEY, Kentucky
ELIJAH CRANE, Arizona                DELIA C. RAMIREZ, Illinois
KEITH SELF, Texas                    GREG LANDSMAN, Ohio
JENNIFER A. KIGGANS, Virginia        NIKKI BUDZINSKI, Illinois

                       Jon Clark, Staff Director
                  Matt Reel, Democratic Staff Director

                SUBCOMMITTEE ON TECHNOLOGY MODERNIZATION

              MATTHEW M. ROSENDALE, SR., Montana, Chairman

NANCY MACE, South Carolina           SHEILA CHERFILUS-MCCORMICK, 
KEITH SELF, Texas                        Florida, Ranking Member
                                     GREG LANDSMAN, Ohio

Pursuant to clause 2(e)(4) of Rule XI of the Rules of the House, public 
hearing records of the Committee on Veterans' Affairs are also 
published in electronic form. The printed hearing record remains the 
official version. Because electronic submissions are used to prepare 
both printed and electronic versions of the hearing record, the process 
of converting between various electronic formats may introduce 
unintentional errors or omissions. Such occurrences are inherent in the 
current publication process and should diminish as the process is 
further refined.

                         C  O  N  T  E  N  T  S

                              ----------                              

                         TUESDAY, MARCH 7, 2023

                                                                   Page

                           OPENING STATEMENTS

The Honorable Matthew M. Rosendale, Sr., Chairman................     1
The Honorable Sheila Cherfilus-McCormick, Ranking Member.........     2

                               WITNESSES

Mr. Daniel McCune, Executive Director of Software Product 
  Management, Office of Information & Technology, U.S. Department 
  of Veterans Affairs............................................     3

        Accompanied by:

    Mr. Charles Hume, Chief Informatics Officer, Veterans Health 
        Administration, U.S. Department of Veterans Affairs

    Mr. Thomas O'Toole, M.D., Deputy Assistant Undersecretary for 
        Health for Clinical Services, Veterans Health 
        Administration, U.S. Department of Veterans Affairs

    Ms. Zhuchun "Emily" Qiu, Director of Health Informatics, 
        Office of Information & Technology, U.S. Department of 
        Veterans Affairs

    Mr. Michael Giurbino, Director of Health Infrastructure and 
        Systems Management, Office of Information & Technology, 
        U.S. Department of Veterans Affairs

The Honorable Roger Baker, Former Assistant Secretary for 
  Information & Technology, U.S. Department of Veterans Affairs..    19

The Honorable James Gfrerer, Former Assistant Secretary for 
  Information & Technology, U.S. Department of Veterans Affairs..    21

Mr. Peter Levin, Co-founder and CEO, Amida Technology Solutions..    22

                                APPENDIX
                    Prepared Statements Of Witnesses

Mr. Daniel McCune Prepared Statement.............................    41
The Honorable Roger Baker Prepared Statement.....................    43
The Honorable James Gfrerer Prepared Statement...................    45
Mr. Peter Levin Prepared Statement...............................    46

                       Submission For The Record

Office of Inspector General Letter to The Subcommittee on 
  Technology Modernization.......................................    49

 
                      VETERANS HEALTH INFORMATION
                         SYSTEMS AND TECHNOLOGY
                          ARCHITECTURE (VISTA)

                              ----------                              


                         TUESDAY, MARCH 7, 2023

             U.S. House of Representatives,
          Subcommittee on Technology Modernization,
                            Committee on Veterans' Affairs,
                                                   Washington, D.C.
    The subcommittee met, pursuant to notice, at 2:58 p.m., in 
room 390, Cannon House Office Building, Hon. Matthew M. 
Rosendale, Sr. (chairman of the subcommittee), presiding.
    Present: Representatives Rosendale, Self, Cherfilus-
McCormick, and Landsman.

      OPENING STATEMENT OF MATTHEW M. ROSENDALE, CHAIRMAN

    Mr. Rosendale. This meeting will come to order.
    I am glad to be here with Ranking Member Cherfilus-
McCormick to continue our work.
    I first want to thank Representative Frank Mrvan for his 
leadership during the last Congress. I really did enjoy working 
with him. As I have stated and he is stated, this is one of the 
most bipartisan committees that we found, that there was a lot 
of joint work together, and we are hoping to be able to do the 
same thing as we move forward.
    We did, however, expose serious mismanagement in several of 
the largest VA information technology systems. Now my goal is 
to motivate the VA to get these efforts back on track or put 
better strategies in place when that is not possible.
    These IT systems exist to make the delivery of care and 
benefits to our veterans easier to schedule, easier to access, 
and easier to track, thereby keeping our veterans healthier and 
happier. Their purpose is not to churn out cushy contracts for 
technology companies. Unfortunately, the Beltway and Silicon 
Valley need to be reminded of that constantly.
    We are here today to discuss Veterans Health Information 
Systems and Technology Architecture (VistA), the platform that 
97 percent of the Veterans Health Administration relies on to 
care for our veterans.
    It is no secret that VistA was originally developed 40 
years ago and Computerized Patient Record System (CPRS) was 
introduced 25 years ago. It is also no secret that VistA is a 
vast collection of hundreds of systems that do everything from 
prescriptions to staffing, to accounting, to tracking car 
accidents.
    It is less well understood where the replacement and 
modernization efforts of 10 and 20 years ago left VistA. When 
the Cerner project began, some in the VA thought they could 
stop paying attention to VistA. Some in Congress thought so 
too. There was a discussion of a pivot plan, which initially 
seemed to be a plan to mothball VistA. As the Cerner completion 
schedule slipped farther and farther into the future, the pivot 
plan faded.
    The reality is, regardless of whether the Oracle Cerner 
implementation can be accomplished and regardless of how we 
feel about that, the VA will probably continue to rely on VistA 
for at least another decade, and some of the elements of VistA 
will probably never go away because no replacement even exists.
    Medical centers all over the country and the veterans they 
serve cannot be left in limbo. Let me say this as plainly as I 
can: Running VistA into the ground would be a disastrous 
mistake. It must be maintained.
    Within the technical constraints that exist, VA should be 
identifying the key areas of VistA that need to be modernized 
and are feasible to undertake. That is already happening in a 
few cases, and I want to explore them this afternoon.
    Let me be clear: There will be no more blank checks and low 
expectations for Electronic Health Record (EHR) projects. VistA 
and Oracle Cerner and any other EHR will be judged by the same 
standards, and those standards are: patient safety, 
reliability, user satisfaction, and cost. It is not 
complicated.
    We have two esteemed panels of experts here with us today 
to do just that, and I want to welcome our witnesses. I 
appreciate you being here.
    With that, I would yield to Ranking Member Cherfilus-
McCormick for her opening statement. Thank you.

OPENING STATEMENT OF SHEILA CHERFILUS-MCCORMICK, RANKING MEMBER

    Mrs. Cherfilus-McCormick. Thank you, Mr. Chairman.
    I am honored to serve as the ranking member on the 
Technology Modernization Subcommittee. Before coming to 
Congress, I was a healthcare executive for many years. I know 
firsthand the critical role that technology plays in the 
provision of healthcare and fully understand that just about 
everything the VA does relies on computers and other 
technologies.
    I would like to welcome our witnesses today and thank you 
for coming to discuss VA's long-serving electronic records, 
VistA.
    While we all know that the VA is in the midst of a major 
modernization effort to replace VistA, it is also important to 
note that the system will be around long as long as it needs to 
be--excuse me--least as long as it needs to be, as long as the 
modernization program takes.
    It is imperative that the VA not only maintains VistA until 
the replacement system is fully deployed but also keeps up the 
healthcare innovation as best as the system allows. Veterans 
who rely on the VA for their healthcare deserve no less.
    There are functions of VistA that are not related to the 
EHR and will likely exist long after the EHR is replaced. To be 
clear, we realize that VistA is more than EHR, but my goal 
today is to focus on the EHR capabilities.
    VA has spent almost 20 years and, according to the 
Controller General, over $1.7 billion in failed attempts to 
modernize VistA.
    Numerous experts told this committee that VistA is not a 
viable long-term solution and must be replaced. There are 
concerns about the cybersecurity risk, data management, and a 
code base considered to be obsolete by many. Also, the 130-plus 
different instances of VistA do not reliably communicate with 
each other, let alone the Department of Defense (DOD) or 
community care providers.
    We must do better.
    Before I conclude, I would like also to congratulate my 
colleagues Mr. Rosendale on being selected as the chairman for 
the subcommittee.
    I look forward to working with you this Congress to ensure 
the VA has the resources and technology it needs to provide the 
world-class healthcare and benefits that our Nation's veterans 
have earned and so richly deserve.
    I yield back the balance of my time.
    Mr. Rosendale. Thank you very much.
    I would like to now introduce the witnesses on both panels.
    First, from the Department of Veterans Affairs, we have Mr. 
Daniel McCune, Mr. Michael Giurbino, and Ms. Emily Qiu from the 
Office of Information and Technology. We also have Mr. Charles 
Hume and Dr. Thomas O'Toole from the Veterans Health 
Administration.
    Thank you all for being with us today.
    On our second panel, we have a distinguished group who led 
VA information technology during the Obama and Trump 
administrations: Mr. Roger Baker served as the Assistant 
Secretary for Information Technology from 2009 to 2013; Mr. 
James Gfrerer served in the same position from 2018 to 2021; 
and Mr. Peter Levin, who was the VA Chief Technology Officer 
from 2009 to 2013.
    I ask the VA's first witnesses on our first panel to please 
stand and raise your right hands.
    [Witnesses sworn.]
    Mr. Rosendale. Thank you.
    Let the record reflect that all the witnesses answered in 
the affirmative.
    Thank you.
    Mr. McCune, you are now recognized for 5 minutes to deliver 
your opening statement.

                   STATEMENT OF DANIEL MCCUNE

    Mr. McCune. Good afternoon, Chairman Rosendale, Ranking 
Member Cherfilus-McCormick, and distinguished members of the 
panel. Thank you for the opportunity to testify today about the 
Department of Veterans Affairs' Veterans Health Information 
Systems and Technology Architecture, or VistA.
    I am accompanied by Chuck Hume, Chief Informatics Officer 
in Veterans Health Administration (VHA); Dr. Thomas O'Toole, 
Deputy Assistant Under Secretary for Health; Mike Giurbino, 
Director of Health Infrastructure and Systems Management in IT; 
and Emily Qiu, Director of Health Informatics in IT.
    Veterans are at the center of everything we do, and VA is 
committed to providing exceptional care, services, and a 
seamless, unified experience to the veteran. To fulfill this 
mission, VA clinicians must have modern tools to ensure best 
outcomes for our veterans.
    Office of Information and Technology (OIT) collaborates 
with VHA and other VA offices to achieve this mission through 
delivery of state-of-the-art technology, including a modernized 
electronic health record, or EHR.
    VistA provides an EHR for veteran care and services, 
supporting over 150 applications in more than 1,500 VA 
facilities. There are 133 instances of VistA nationwide that 
share common functionality but have data and workflow tailored 
to the needs of each medical center and patient population.
    VistA has served VA and veterans for over 40 years, and we 
are aware of its limitations. It does not have modern 
capabilities like artificial intelligence, machine learning, 
mobile and web access, and the capabilities providers and 
veterans expect and deserve from a modern, cloud-native EHR.
    In 1997, VA implemented a new graphical user interface on 
top of VistA called Computerized Patient Record System, or 
CPRS. CPRS is a Windows desktop application that provides an 
improved, yet dated, user interface for accessing VistA data.
    VistA itself is written in an old programming language 
called Massachusetts General Hospital Utility Multi Programming 
System (MUMPS). There are few MUMPS programmers today. MUMPS is 
not taught in computer science classes, and the pool of MUMPS 
programmers shrinks every year as they retire.
    VA is fortunate to have dedicated MUMPS programmers 
supporting VistA. They understand millions of lines of code 
developed over 40 years, and they also understand VA clinical 
business processes. They are committed to enabling clinicians 
and supporting veteran outcomes, and we have been able to 
retain them and their knowledge much longer than a typical 
workforce. However, approximately 70 percent of our MUMPS 
developers today are retirement-eligible, and we have few 
options to hire or contract additional ones.
    VistA is a member of VA's expansive and complex ecosystem 
of software and infrastructure. The size and complexity of that 
technology ecosystem has nearly doubled in the last 5 years, 
and most of that growth has been in modern, cloud-native 
applications.
    MUMPS programmers are increasingly challenged keeping VistA 
integrated in a growing ecosystem that is architected very 
differently from the system designed 40 years ago. While 
technology is a challenge, so also are the dated skills of our 
VistA programmers. That challenge compounds every year.
    In May 2018, after nearly a year of analysis, VA decided to 
replace VistA with a Commercial Off The Shelf (COTS) solution, 
Cerner Millennium. However, VA will continue to use VistA until 
Cerner Millennium is fully implemented. During this time, it is 
essential to maintain and enhance VistA to preserve the 
uninterrupted care and continually improve the veteran service.
    Some of the key recent VistA enhancements include: 
standardizing VistA code, we call this our national gold 
version; move to cloud: 20 instances of VistA have been moved 
to the cloud, with an additional 54 planned this year; 
improving access to federated health data by implementing an 
Application Programming Interface (API) gateway.
    Currently, there is no cost reduction tied to the new EHR 
solution. The cost to support VistA will increase as we develop 
new capabilities and interfaces, support congressional 
mandates, and migrate to cloud. In essence, we are supporting 
two EHR systems simultaneously until the Cerner implementation 
is complete.
    In the interim, VistA remains our authoritative source of 
veteran data.
    In summary, VistA is an old technology, ill-suited for the 
modern digital age. Modernization would require VistA to be 
rewritten almost from scratch at a great cost and great risk.
    Chairman Rosendale, Ranking Member Cherfilus-McCormick, and 
members of the subcommittee, thank you for the opportunity to 
appear before you today to discuss OIT's improvement of VistA. 
This concludes my testimony, and I look forward to answering 
your questions.

    [The Prepared Statement Of Daniel Mccune Appears In The 
Appendix]

    Mr. Rosendale. Thank you, Mr. McCune. I appreciate that.
    We are going to go straight to questioning.
    I mentioned in my remarks that I judge any EHR or any 
system that supports healthcare according to patient safety, 
reliability, user satisfaction, and cost. Does the VA 
typically--or, currently use any other criteria for 
establishing what they would like to have?
    Mr. McCune. That is a good start, sir. I think we do look 
at reliability and availability of the system too. Access to 
our critical health information systems is very important to 
our veterans and to our clinicians.
    Mr. Rosendale. Do you have any other criteria that you are 
using, because reliability and customer satisfaction, the cost, 
these are all things that we are already looking at.
    Mr. McCune. Yes, sir.
    Mr. Rosendale. What would they be? Exactly why would you be 
using these other criteria?
    Mr. McCune. Sir, if--thank you for the question. If we are 
looking at how do we gauge success of a large-scale 
implementation--is that the question?
    Mr. Rosendale. Sure.
    Mr. McCune. Okay. There is a number of factors that we use. 
The ones you mention--scope, schedule, and cost--certainly are 
high on the list. Customer satisfaction, also high on the list. 
Those are ones that we look at.
    We look at the technical viability and how it fits in our 
ecosystem. I mentioned our large ecosystem.
    Those are all elements that we take into account.
    Mr. Rosendale. Do you take into account the ratings or the 
consideration from the employees themselves, the ability for 
them to work with the system and the problems that it may not 
cause or the efficiency that it delivers to them?
    Mr. McCune. Yes, sir. We consider both the veteran 
experience and the user experience.
    Mr. Rosendale. How exactly are you measuring that?
    Mr. McCune. We have a number of different ways to measure 
that, sir. One of the key ways we look at that is both the 
accessibility and the performance of that system. We look at 
numbers--how many clicks does it take, how many screens do they 
need to see, how performant is that system. Those are all 
measures that we use to gauge user acceptance.
    Mr. Rosendale. Are you going back and having a poll 
conducted with the actual users and the patients to see what 
their experience has been?
    Mr. McCune. Yes, sir. I think that is an excellent question 
for our VHA partners. I would ask Mr. Hume to answer that 
question.
    Mr. Hume. Sir, last year, we did the initial survey of 
users in Veterans Integrated Services Network (VISN) 10 and 20. 
That is where the Cerner product is currently deployed. We used 
standard survey questions developed by KLAS. I do not recall 
what that acronym is for, but it is an industry leader who 
surveys the electronic health record systems across the country 
and user experience with that. We got some baseline data from 
that survey last year.
    Mr. Rosendale. Okay.
    Where is the survey? That is what I am looking for, right 
there.
    Okay. I happen to have a copy of that survey, and I have 
got some serious concerns about this, okay? After being 4 years 
in service, the VA's version of Cerner ranked dead-last in the 
survey that I am looking at.
    Okay. After 4 years, according to this survey, 78 percent 
of the users--so these are the docs, right--78 percent of the 
users do not feel that it helps them deliver high-quality care. 
Seventy-eight percent. We are talking about almost 8 out of 10.
    On the other hand, 64 percent said that VistA enabled them 
to deliver high-quality care. It is almost a complete flip-flop 
of that.
    Now, we are talking about after billions of dollars have 
been expended, we are talking about several years to try and 
work through this system, and we are talking about at some of 
the smallest facilities in the country, not the more complex 
centers that the VA is even responsible for.
    My question would be: Mr. McCune, does anything in the KLAS 
report surprise you?
    Mr. McCune. Sir, I think--thank you for the question. I 
would say, no, sir. I think we are getting a lot of signal 
around that.
    What I would say is, with VistA, we have a relatively 
stable system, one that is been in production for 40 years. Our 
clinicians, our users are very, very familiar with that system.
    What we also have happening is a brand-new system, the 
Cerner system. I think there is some element of change 
management there, there is some element of newness that has to 
be considered there.
    We are aware of those numbers, sir, and we are tracking 
those.
    Mr. Rosendale. Even though VistA is this old, why do most 
of the employees like it so much, Mr. McCune?
    Mr. McCune. Sir, I would defer to our VHA partners on that 
one.
    Mr. Hume.
    Mr. Hume. I will invite Dr. O'Toole to answer as a user of 
CPRS.
    Dr. O'Toole. Thank you, Congressman.
    I think the issue with CPRS, and having been using it 
myself for the past 17 years, is that it accomplishes what we 
need. It does represent high reliability.
    But I would be neglectful to not say it does not have its 
own problems, and it does not provide a lot of the 
functionality that we see in different EHRs and more modern 
EHRs.
    There is a muscle memory associated with using it for quite 
some time that I believe providers are comfortable with. They 
know how to use it, they know how to navigate the system, and 
it has worked well for us.
    Mr. Rosendale. Mrs. Cherfilus-McCormick, do you have 
questions?
    Mrs. Cherfilus-McCormick. Yes. Thank you.
    Researchers have noted current challenges with VistA and 
how it negatively impacts healthcare delivery and operations, 
to include quality healthcare delivery, security, care 
standardization, lack of system standardization, data 
standardization, access controls, and efficient business 
operations.
    That is no small list of problems. Would you agree that 
those areas are of concern?
    Mr. McCune. Yes, ma'am, I would agree with you. There is a 
long list with our legacy technology and our legacy VistA 
platform.
    Mrs. Cherfilus-McCormick. On the major issue of lack of 
standardization, I realize that this is partly the fault of the 
VHA and their lack of prioritization of creating a standard 
healthcare record.
    How difficult would it be to standardize the health records 
across the separate instances of VistA across the country, and 
how long would it take?
    Mr. McCune. Thank you for that question, ma'am.
    That is a long, complex question, and we have not done the 
analysis on that. We have tried over the last 40 years many 
times to tackle that problem. If you would like, we can take--
can do the analysis and come back with a plan.
    Mrs. Cherfilus-McCormick. Just for information, how many 
instances of VistA are there? Is it 130 instances?
    Mr. McCune. It is--technically, it is 133.
    Mrs. Cherfilus-McCormick.--thirty-three.
    In other situations, have we seen any organization try to 
actually standardize? What would be that average time? Have we 
seen it in 50? Sixty? What would that standard time be, if we 
do not know what 133 variations would be?
    Mr. McCune. Yes. Thank you for that question.
    Ma'am, I do not think we have a good comparable. I think 
that is what you are asking: Do we have a good comparable? I do 
not think we have a comparable. Nothing is the size and scale 
of our EHR. That is one of our largest--it is our largest 
technology system.
    Mrs. Cherfilus-McCormick. There has been some discussion in 
Congress and elsewhere to make another attempt to modernize 
VistA and abandon the current Oracle Cerner system.
    I am not here to say the Oracle Cerner approach in 
Electronic Health Record Modernization (EHRM) is going well, 
but I am not sure returning to VistA is the correct either. We 
have heard from inside VA that VA does not have the expertise 
or the desire to go the pathway of modernizing VistA, which 
would itself take several billion dollars, other large IT 
contracts, and large management contracts such as the VA has 
today.
    Is that accurate?
    Mr. McCune. Thank you for that question. Ma'am, I think, if 
I heard you correctly, that is a fair summarization.
    Mrs. Cherfilus-McCormick. What is precluding VA from taking 
a more measured approach?
    Mr. McCune. Ma'am, as we look to our limited resources, 
right, our resource constraints and how we focus those, we have 
them currently 100 percent focused on the current EHR 
implementation. We do not have extra people working on a plan 
B.
    We are fully committed to this. Until such a time as we 
need to change course, all of our resources, both our people 
and our funding, are allocated toward the successful Cerner 
implementation.
    Mrs. Cherfilus-McCormick. My largest concern with the 
current EHRM program is that staff do not feel like their 
concerns are being taken into account, which is resulting in 
potential patient harm.
    Is there not a better way to provide a more--a better roll-
out of the Oracle Cerner system while also providing some 
clarity on what will and will not be available for their use in 
VistA?
    Mr. McCune. Thank you for that question.
    Ma'am, I think what you have today are the people 
representing VistA. I do not think we are the right people to 
represent the EHRM solution or the challenges that they are 
facing. I would defer that question. We can take that for the 
record.
    Mrs. Cherfilus-McCormick. Okay. Thank you. I yield back.
    Mr. Rosendale. Thank you very much, Representative.
    I would now go to my friend from Texas, Representative 
Self.
    Mr. Self. Thank you, Mr. Chairman.
    I want to explore the cloud migration.
    First of all, I am not an IT person. I have looked for it 
in your testimony. What is the Inquiry Routing and Information 
System (IRIS) for Health 2022.1? What does that mean?
    Mr. McCune. Sir, I think I--can you repeat your question?
    Mr. Self. What is your reference to IRIS, I-R-I-S, for 
Health 2022? It is part of your cloud migration testimony.
    Mr. McCune. Yes, sir.
    Mr. Giurbino, can you answer that question?
    Mr. Giurbino. Yes, sir.
    It is our data base platform that we are upgrading to, that 
brings us new capabilities.
    Mr. Self. You are upgrading VistA, the platform as well, 
plus migrating it to the cloud.
    How much does it cost to move it to the cloud? When will 
you complete that? How important is that for the future?
    Mr. McCune. Yes, sir. Thank you for that question.
    Cloud migration is an important component in maintaining 
and enhancing VistA.
    I would say that the version of cloud migration that is 
taking place today is what we call a ``lift and shift,'' 
meaning we are not changing VistA; we are simply taking it out 
of our data center and moving it into a virtual cloud 
environment. That is what we call a ``lift and shift.''
    We are in midstream there. About 20 instances today have 
been migrated. An additional 54 are planned this year.
    That is not modernization. Modernization, making it a 
cloud-native application, would involve rewriting that code.
    I think you asked about cost, sir. I think it cost us 
around $70,000 per instance to move it to the cloud.
    Mr. Self. For 130, rough?
    Mr. McCune. Yes, sir. Roughly, it would be about $70,000 
per instance.
    Mr. Self. Is Cerner going to be on the cloud?
    Mr. McCune. Yes, sir, Cerner is in the cloud.
    Mr. Self. Okay.
    Let us move to the community care providers. How many 
providers do you refer veterans to?
    Mr. McCune. Sir, I will hand that off to Mr. Hume.
    Mr. Hume. Sir, I think we are going to have to take that 
one for the record.
    We do health information exchanges with 130--excuse me--
with members of the eHealth Exchange, the national eHealth 
Exchange, as well as CommonWell and Carequality. We are moving 
toward Carequality.
    We do information exchanges with partners who have joined 
the health information exchange, but that is not the universe 
of providers that we refer patients to.
    Mr. Self. Then tell me how VistA integrates with the 
70,000. What is the integration there? Is there any?
    Mr. McCune. Sir, 70--I am sorry. I do not understand your 
question, sir. Can you repeat that?
    Mr. Self. For your community care providers that you 
coordinate with, how does VistA integrate with them? Then I 
would like to compare Cerner to that as well.
    Mr. McCune. Thank you, sir. I am going to ask my partner 
Ms. Qiu to answer that question.
    What I can tell you in broad statements is: The integration 
that VA does--or VistA does with our community care providers 
is not nearly as comprehensive or as native as what we have 
with Cerner. We are doing some things to do that integration.
    Ms. Qiu, can you answer that question?
    Ms. Qiu. Sure. Thank you, Mr. McCune.
    To communicate with our community care providers, we are 
implementing a solution called VDIF. It stands for Veterans 
Data Integration and Federation Platform.
    This platform is connecting to Cerner's Joint Health 
Information Exchange (JHIE). At request, VDIF would respond 
with the longitudinal patient record to Cerner JHIE, and then 
Cerner JHIE would relay that message back to the community care 
providers.
    Mr. Self. You just addressed Cerner, correct?
    Ms. Qiu. That is correct. We use Cerner JHIE to connect to 
two networks. One is CommonWell and the other is eHealth 
Exchange Network, which Cerner JHIE connects to.
    Mr. Self. You already have a complex system of both VistA 
and Cerner in your coordination with your community care 
providers. Is that correct? Am I hearing that right?
    Ms. Qiu. To make the community care communication work, we 
are using both the VDIF solution and Cerner JHIE solution to 
make it work.
    Mr. Self. Thank you, Mr. Chairman. Sounds complex to me.
    Mr. Rosendale. Thank you very much.
    I would now yield to the gentleman from Ohio, Mr. Landsman.
    Mr. Landsman. Thank you, Mr. Chair. I appreciate the time.
    Thank you all for your service and for being here today.
    A few questions, three. Some of this I suspect you can 
speak to today, and some of it you will have to, you know, 
circle back.
    One has to do with what we have learned from the five sites 
where this has been implemented: you know, Columbus is close to 
where I serve, southwest Ohio, Cincinnati; but Walla Walla, 
Roseburg, Spokane, White City.
    What are the high-level takeaways from these pilots? I 
mean, what would be most helpful for us to know in terms of 
what you have seen, both the good and the bad, in these five 
sites?
    Mr. McCune. Thank you, Congressman.
    I do want to circle back, if I can, for----
    Mr. Landsman. Sure.
    Mr. McCune [continuing]. just a second on the cloud 
question. I think the question was, is Cerner in the cloud?
    While the target is to move Cerner--or to have Cerner in 
the cloud, it is not currently in the cloud today. That is a 
work in progress. I just wanted to correct that.
    To your question, sir, I think I am going to defer that to 
Dr. O'Toole.
    Mr. Self. Okay.
    Dr. O'Toole. Thank you, sir.
    We have learned a lot from the five deployments. I think it 
has exposed a lot of challenges in both product and product 
development, refinement; our staff training; and stresses to 
the system. It is a learning curve.
    Some provider groups have fared better than others in terms 
of their adoption. I think that the decisions to date by the 
agency to put a hold on the Boise deployment, to put a hold on 
the Ann Arbor deployment, reflect the fact that we are actively 
working to make improvements both to what we do internally, to 
what is expected of the product, to what is expected of the 
deployments, before we go live with additional sites.
    Mr. Landsman. Yes. I mean, I think what we--that is 
helpful. Perhaps for the record--so this is a request--to just 
better understand what the high-level, you know, learnings, 
takeaways have been, both positive and negative. I mean, I know 
we are struggling on a whole host of levels, but what is 
promising? What do you feel good about, or what do you see as 
big opportunities?
    I am assuming that is part of what we will see more of 
after the pause, right? I mean, once--and we do not have to get 
into--if there is high-level stuff you want to share, please 
do, but I do think it would be helpful for us to better 
understand those big takeaways.
    Dr. O'Toole. Thank you. I definitely appreciate the 
importance of the question.
    There will be a report that is still in preparation related 
to the sprint efforts----
    Mr. Landsman. Okay.
    Dr. O'Toole.--that should be released shortly as it goes 
through. I think that will help define many of the criteria 
that are expected to be in place for future go-lives that I 
think would answer some of those questions.
    I would request that we take for the record, you know, 
pending that.
    Mr. Landsman. Then the second question has to do with 
leadership changes at the Integration Office. As these changes 
have taken place, what are the gains that--I mean, what are we 
not losing I think would be helpful to know.
    How have you all managed or thought about that, the change 
in leadership, as it relates to ensuring that what needs to be 
learned is learned and what needs to be changed is changed and 
so on and so forth?
    I mean, that is a huge question I have, which is, when you 
make a big leadership change, when you are dealing with this 
kind of complexity, how are you handling that?
    Mr. Hume. Well, I would just acknowledge that Dr. Neil 
Evans is taking over the office as the Acting Director, 
Executive Director, and we have great confidence in him. He has 
been a VA clinician for many years, and he is talented in this 
space, and we are looking forward to his leadership.
    Mr. Landsman. I think similar to the previous question, the 
gains that were--and improvements, the opportunities that his 
predecessor, you know, experienced, you know, making sure those 
are not lost. I mean, how do we--and things that, obviously, 
were not going well. I think having a better understanding--and 
maybe that will be in the report or----
    Mr. Hume. In that position, there are three primary 
deputies, and they are all staying in place. There is good 
continuity at the next level down from the Director.
    Mr. Landsman. Then, finally, this has to do with the 
contracting. I mean, so much of this gets outsourced to, you 
know, private companies. I think it would be helpful to this 
subcommittee to better understand what you are concerned about 
in terms of that contracting process.
    What is being communicated effectively when you are 
securing support from outside vendors? You know--and this gets 
to some of the questions around the requirements, you know, 
that may or may not have been articulated when you went out, 
you know, to get help.
    You know, it would help me to better understand how you all 
are approaching that moving forward, because, clearly, you 
know, you are going to--it requires a tremendous amount of 
support, and the Department's ability to get the right kind of 
help, having articulated the right set of requirements, is 
going to be very, very important.
    I am wondering if someone can speak to that and/or circle 
back.
    Mr. McCune. Thank you for that question, sir. I think part 
of that answer is to--and we will take back. I will address 
part of that question, which is us working with contractors. I 
think your question is probably around accountability.
    Mr. Landsman. Yes.
    Mr. McCune. Working with vendors is a normal part of what 
we do in VA, and particularly in my area, which is systems 
development. Almost 85 percent of our workforce today is 
contracted, and so that is really a partnership between the 
govvies and the contractors.
    Mr. Landsman. Thank you.
    Mr. Rosendale. Thank you, Mr. Representative.
    Dr. O'Toole, let us review the information that we have 
about VistA. The subcommittee submitted questions in advance of 
this hearing, and the VA has answered them.
    First, patient safety, which is a priority for everybody in 
this room. The VA provides statistics on the number of patient 
safety reports related to the system. How many patient safety 
reports were there in each of the last 3 years across the 166 
medical centers and their clinics using VistA?
    Dr. O'Toole. Sir, I am looking through my notes here, in 
terms of what we submitted in response to that question.
    Mr. Rosendale. Okay.
    Dr. O'Toole. I believe the data, which reflected two 
different modes of presenting, shows a significant increase. It 
is not broken down by medical center.
    In terms of what are referred to as Joint Patient Safety 
Reports, or JPSRs, which reflects a concern raised by the 
field, not necessarily a validated concern, not necessarily 
something that was reflected in a safety event or patient 
harm--and I think that is an important----
    Mr. Rosendale. Let me tell you the numbers that I have, and 
you----
    Dr. O'Toole. Yes.
    Mr. Rosendale [continuing]. can tell me if this is accurate 
or not.
    Through the 166 medical centers and their clinics using 
VistA, we have in 2020, 12,644; in 21, we did have an increase, 
it was 14,637; but then in Fiscal Year 22, it went all the way 
down to 9,211.
    Dr. O'Toole, the information we have from Spokane for the 2 
years after Cerner went live is 1,033 patient safety reports. 
This is something that this committee has complained about for 
quite some time, extensively, in a bipartisan manner. That is 
over 500 reports per year from 1 hospital using Cerner--1 
hospital--compared to an average of 55 reports annually from 
the VistA hospitals.
    How do you explain this huge difference? In short, why has 
the Mann-Grandstaff VA Hospital become the most dangerous VA 
hospital in the country?
    Dr. O'Toole. Well, thank you for the question, sir.
    Again, I am not going to try to minimize, by any stretch, 
the concerns that staff raised at Mann-Grandstaff and continue 
to raise at Mann-Grandstaff regarding safety concerns. I think 
it has been well-documented and well-noted, challenges that we 
have had with the deployment of the Cerner product, the Cerner 
Millennial product, at Mann-Grandstaff related to a lot of 
issues, not the least of which is deployment during the 
pandemic, deployment with significant concerns about product 
readiness, training, and so on.
    We are striving to be a high-reliability organization. We 
continue to encourage these reports to be generated so they can 
be investigated and acted on. To the extent that we continue to 
act on them, as reflected by the pauses that have occurred in 
our deployment of future sites while actively deploying tiger 
teams and others to address those concerns, I think has, you 
know, been well-noted.
    You know, not to make any excuses about the concerns at 
hand.
    I will say that there have been improvements to the Cerner 
product over time, and we have made several improvements in the 
care delivery, in the workstream----
    Mr. Rosendale. Can I just cut to the chase and say, would 
you attribute this increase to the Cerner system, Dr. O'Toole?
    Dr. O'Toole. Yes.
    Mr. Rosendale. Thank you.
    Dr. O'Toole, I asked the Office of Inspector General to 
identify the reports related to VistA and patient safety in 
preparation of this hearing. They looked back 6 years and found 
five reports.
    I am going to enter this letter from the Office of 
Inspector General (OIG) into the record.
    Mr. Rosendale. These all involve automated view alerts, 
which let providers know test results are available, and the 
lack of ability to track who observes and acts on those alerts.
    What component of VistA generates and tracks these alerts? 
How are you addressing this issue?
    Mr. Hume. Well, I believe you are talking about reminder 
alerts and reminders that the system generates usually as a 
result of the configuration of--or mandates to perform certain 
tasks, clinical tasks--Dr. O'Toole would be better to answer 
that--perform certain tasks at certain times, and the providers 
are reminded through those alerts to perform those tasks.
    Dr. O'Toole. For clarification, is that what you are 
referring to, Congressman?
    Mr. Rosendale. Yes.
    Dr. O'Toole. Okay.
    Yes, so the clinical reminder and clinical alert system 
within VA is something that has been used for decades now as a 
means of reminding providers both for specific screening 
questions, view alerts for patients who may be at risk for 
certain events, as well as reminders for different clinical 
activities.
    It is something that is currently tracked and actively 
tracked and fed back to the Department through our data and 
analytics group.
    Mr. Rosendale. Again, the question was: What component of 
VistA generates and tracks these alerts? How are you addressing 
this issue that they are not being resolved?
    Dr. O'Toole. Well, sir, I think that there are several 
concerns with that, not the least of which is that there are 
probably too many alerts and not all of them are relevant, and 
there is an issue of alert fatigue, which we are actively 
looking at in terms of trying to streamline some of those 
alerts so that the most pertinent ones, the ones that are most 
timely to the clinical event and patient care issue and where 
the risk is highest are those that are prioritized and others 
that may be less important can be sunsetted. That is a current 
and active issue that is going on.
    All of the clinical stations, though, do receive reports on 
compliance with that and are being tracked, both nationally and 
locally, in terms of their performance in that space.
    Mr. Rosendale. Representative Cherfilus-McCormick, do you 
have more questions?
    Mrs. Cherfilus-McCormick. Thank you, Mr. Chair.
    If the VA is itself admitting that it does not want and 
does not have the capacity to manage a VistA modernization, why 
would we, Congress, force them to?
    Mr. McCune. Thank you for that question, ma'am.
    I do not think we are asking you to force us to do that, 
right? I think what we have today and what is before Congress 
is a difficult choice. We have got a system implementation 
underway today, and we have an aging legacy system that we have 
struggled multiple times in the past to modernize. It is not an 
easy choice, sir--or ma'am. It is one that we are continually 
looking at.
    Mrs. Cherfilus-McCormick. Previous testimony indicates the 
VA spends about $800 million a year to maintain, improve, and 
stabilize VistA. Is that accurate?
    Mr. McCune. Yes, ma'am.
    Mrs. Cherfilus-McCormick. Is it accurate that VA continues 
to keep VistA healthy and plans to until the last proposed EHRM 
deployment is finished? In other words, the VA is not walking 
away from VistA today, correct?
    Mr. McCune. That is correct, ma'am.
    Mrs. Cherfilus-McCormick. Mr. Chair, I yield back.
    Mr. Rosendale. Representative Self?
    Mr. Self. Thank you, Mr. Chairman.
    You have testified that you have delayed Boise and Ann 
Arbor. Given that you restart your Cerner roll-out, when do 
you--when would you anticipate the last use of VistA?
    Mr. McCune. Thank you for that question, sir.
    I think that timeline is a little uncertain. As we started 
the Cerner roll-out, we were planning for 10 years. VistA is 
our interim solution. It is an interim solution for an 
indefinite amount of time. Five to 10 years is the time window 
we are looking at now, but that may extend.
    Mr. Self. From what I am seeing, let us take the 10 years 
out, with the delays. What is VistA going to look like in 10 
years?
    Let us go back to your MUMPS coders. This is not unique, by 
the way. Companies are pulling Common Business Oriented 
Language (COBOL) coders out of retirement. It is not unique 
across the industry. Old codes are--old programs are being 
resurrected because of their simplicity and so forth.
    But--so what is VistA going to look like in 10 years?
    Mr. McCune. Thank you for that question, sir.
    We do have COBOL programmers as well.
    Mr. Self. Okay.
    Mr. McCune. It is hard to tell what VistA is going to look 
like between now and then, for a number of reasons. It is a 
moving target.
    Number one, we are going to need to make changes to VistA 
in order to support the Cerner roll-out, particularly 
integrations.
    Number two, we are going to have emerging requirements from 
our clinicians, and so we are going to need to adapt VistA.
    We are also going to need to adapt VistA to changing 
technologies. You are seeing some of that right now with our 
move to cloud.
    So, predicting 10 years from now, sir, that is difficult to 
do for VistA. All the problems we have talked about--aging 
workforce, a programming language that is not taught today--all 
of those concerns are compounded particularly if you talk about 
a 10-year window.
    Mr. Self. That leads me to the $17 billion for Cerner. You 
said it would be very expensive to rewrite VistA. Do you have 
an estimate on that?
    Mr. McCune. No, sir, we do not. We have not done an 
analysis on a large-scale full modernization of VistA. We 
have--I think the committee is equally versed on the cost 
estimates for Cerner. We have not done that similar kind of 
cost estimate for VistA.
    Mr. Self. I would assume you did that, though, before you 
gave the--before you awarded the Cerner contract, because do 
not you examine all your options?
    Mr. McCune. Yes, sir. We considered all of our options back 
in 2017 and 2018. We had about a year's worth of analysis that 
went into making that decision.
    Any data we had around VistA and its viability for 
modernization is old and dated, and we would have to revisit 
it.
    Mr. Self. I have been handed a piece of paper that looks 
like far more than my $17 billion.
    I would--so you did or did not do that analysis before you 
awarded Cerner?
    Mr. McCune. Sir, thank you for that question.
    We did do the analysis back in 2017 and 18 on all of our 
options, and that is when we chose the Cerner solution.
    Mr. Self. Do you remember what the VistA rewrite was?
    Mr. McCune. No, sir, I do not, but we can get that 
information.
    Mr. Self. I would like to see that.
    Mr. Self. Thank you, Mr. Chairman.
    Mr. Rosendale. The gentleman from Ohio, Mr. Landsman.
    Not there? Okay.
    Representative Cherfilus-McCormick. Go ahead.
    Mrs. Cherfilus-McCormick. Thank you, Mr. Chairman.
    In 2017, Government Accountability Office (GAO) issued a 
report documenting the challenges with VistA's pharmacy system. 
The report was entitled ``Pharmacy System Needs Additional 
Capabilities for Viewing, Exchanging, and Using Data to Better 
Serve Veterans.''
    Many of VA's plans to address the pharmacy system were to 
use the Oracle Cerner pharmacy solution as part of the EHR 
suite. If VA were to abandon the Oracle Cerner product, would 
VA need to go back and fix each of the pharmacy issues outlined 
in that report and others?
    Mr. Hume. Well, many of those findings were addressed--have 
been addressed over the last 5 years, particularly with respect 
to a graphical user interface for pharmacy.
    We would have to take for the record a specific--the 
specific things that were done and were not done out of that 
report.
    Mrs. Cherfilus-McCormick. Do you know how many were already 
fixed versus how many are left?
    Mr. Hume. We would have to take that, ma'am.
    Unless you are able to answer that? No.
    Mrs. Cherfilus-McCormick. Okay. VistA may have less 
problems today, but is part of that because everyone knows how 
to use it and knows its quirks?
    It is like a stable old family car. Everyone knows you need 
to jiggle the clutch when it gets stuck in third gear. Would 
you agree with that statement?
    Dr. O'Toole. Yes. Yes, ma'am.
    Mrs. Cherfilus-McCormick. What patient safety issues comes 
up with VistA? My concern is the lack of coordination for all 
veterans, with a focus on veterans often overlooked and left 
out.
    Dr. O'Toole. There are a lot of safety concerns within 
VistA that, you know, I would be neglectful to say do not 
exist. A lot of them are instance-specific and not necessarily 
generalized to the entire system, I think some of which we 
referenced earlier with the clinical reminders as one element, 
view alerts being another one. Again, it is difficult to kind 
of speak to it in generalities.
    You know, I think we do a better job than most systems in a 
lot of our population health work and a lot of our tracking of 
potentially vulnerable populations. Could it be done better, 
and could there be system improvements either within VistA or 
within other systems to enhance those capabilities, you know, I 
think, are clearly questions that would need to be explored.
    Mrs. Cherfilus-McCormick. Thank you so much for your 
answers.
    Mr. Chairman, I yield back.
    Mr. Rosendale. Thank you very much, Representative.
    Mr. McCune, let us turn to reliability on the system, 
uptime. According to the information provided, VistA's uptime 
in 2022 was 99.9967 percent--that is better than Ivory soap--
and in 2021, it was 99.999964 percent. This is almost perfect. 
That is excellent.
    How in the world is that possible even through upgrades and 
changes?
    Mr. McCune. Thank you for that question, sir.
    We strive to get to four 9's, and with VistA we have been 
able to do that.
    What I would say is that you have new systems. Generally, 
as we release new systems, we have more issues, particularly 
with uptime and reliability, than we do with systems that have 
a long life span.
    With VistA, for instance, we have had decades to work out 
the bugs, fine-tune performance, and that is how we are able to 
get to those high reliability rates.
    Mr. Rosendale. Mr. McCune, is it true that if one VistA 
instance goes offline for whatever reason, some reason or 
another, the other instances may not be affected?
    Mr. McCune. That is correct, sir.
    Mr. Rosendale. Okay.
    This brings me to one of the problems that I have been made 
aware of. How does that compare with the Cerner system? It is 
my understanding that when a system goes down, it is across all 
instances, in many cases, and during upgrade periods.
    Mr. McCune. Sir, I think I can speak to VistA. I do not 
feel comfortable speaking to the Cerner implementation and 
their configuration.
    I will tell you that they are on-prem today and they are 
looking to move to cloud. I would expect different kinds of 
availability once that migration is complete.
    Mr. Rosendale. Okay.
    Mr. McCune, the VA also provides statistics on the 
workforce that maintains VistA. This committee has been hearing 
for years that the employees who are able to write the code are 
retiring--you said it just earlier, as a matter of fact--and 
can not be replaced, but the numbers do not bear that out.
    In 2022, 1,129 full-time equivalents, FTEs, worked on 
VistA, and 10 years ago, it was almost the exact same number at 
1,134. That is only five people different.
    The VA seems to be maintaining a stable workforce even 
after all of the challenges that everybody seems to be facing. 
How are you able to keep those folks on?
    Mr. McCune. Yes. Thank you, sir, for that question.
    We have been fortunate. I think our employees are committed 
to the mission, and they have stayed long beyond what a typical 
workforce would stay.
    That number I mentioned earlier, 70 percent retirement-
eligible, that number has been creeping up year over year. 
Eventually we are going to start to lose them.
    Mr. Rosendale. Eventually. I mean, right now that is the 
definition of a stable workforce. I mean, a 10-year period and 
you have virtually the same number of people there.
    Mr. McCune, the VA provided figures for spending on VistA 
over the last 5 years. In 2022, VA spent, excuse me, $890 
million, including about $70 million in development.
    What factors drive the steady-state cost, and how do you 
determine the amount to allocate development?
    The reason I ask that is because I look at fiscal 18, 19, 
20, 21, 22, and the development modernization enhancement, DME, 
has changed dramatically: $16 million, call it $17 million, in 
18; $25 million in `19--I am sure you have these numbers in 
front of you--$51 million in 20; $104 million in 21; and then 
drops down again to almost $70 million in 22.
    What factors drive this?
    Mr. McCune. Thank you for that question.
    Sir, every year, we go through the budget drill and we set 
priorities for the administrations. What you will see in the 
variability with development and enhancement is a reflection of 
those priorities. Sometimes there is an improvement needed to 
VistA and that rises to the level of funding; other times it 
does not.
    Mr. Rosendale. Okay. I will tell you my concern, and that 
is, when we see that Cerner is now on year 5 of 10 and we were 
told about this roll-out taking place and we look at 18 and we 
see the development modernization enhancement being starved out 
of VistA, I have grave concern when I also hear that we are 
relying so heavily on VistA still, not only for separate 
facilities--166 of the 171 facilities--but for the existing 
facility, even some of the functions we are still relying upon 
VistA, and yet we are basically creating a self-fulfilling 
prophecy by starving that system of the investment that it 
needs so that it can perform the very duties that you are 
calling upon it for.
    I mean, can you respond to that at all?
    Mr. McCune. Yes, sir. I appreciate the concern around 
funding levels, particularly for development of VistA.
    What I will tell you, sir, is that our resources are 
limited. Right now, both our people and our budget are fully 
committed to that Cerner implementation.
    Mr. Rosendale. Mr. McCune, I understand that your funds are 
limited, okay, and that you are budgeting. When I see a system 
that is providing the vast amount of healthcare delivery 
information for you and I see you starving it, and then you 
come in before us and tell us that it is not only being 
utilized for the vast majority of the facilities but the 
facilities that are utilizing Cerner are still depending upon 
this system as well, why would you still starve that system of 
the revenue it needs to provide the technology that is working?
    Mr. McCune. Yes, sir. Most of our funding today around 
VistA is with that integration with Cerner. Again, it is 
limited resources, sir. A lot of funding is being spent on 
migration of data, on building integrations between VistA and 
our legacy systems and the Cerner system.
    Yes, sir, you are seeing a tapering off of development on 
the VistA platform.
    Mr. Rosendale. At the same time, we have got a $1.1 billion 
investment that is being sent to Cerner, again, where we see 
five facilities that they are funding that are not working 
properly, that are still relying on the VistA system.
    Do you think that is a wise allocation of funds, Mr. 
McCune?
    Mr. McCune. Sir, I can speak to the VistA system and our 
funding and allocations there. I would take it back for the 
record and our experts on the EHRM side of the house.
    Mr. Rosendale. Okay.
    Mr. Rosendale. Mrs. Cherfilus-McCormick, I am sorry, I went 
way over. Do you have any additional questions for this round?
    Mrs. Cherfilus-McCormick. No.
    Mr. Rosendale. Mr. Self, do you have any additional 
questions?
    Mr. Self. Thank you, Mr. Chairman. I just want to follow up 
on your point there.
    Will there be an end to VistA? Will the integration, will 
it be required well beyond Cerner roll-out?
    I think I am hearing there may not be an actual end to 
VistA. Am I hearing that correctly?
    Mr. McCune. Thank you for that question, sir.
    I think there is a couple of facets to that question. One 
is what is going to be replaced by Cerner. Of the, I think, 150 
modules, all but 5 of those are targeted to be replaced by 
Cerner, so very little would be left over from a VistA 
perspective.
    On the VistA side of the house, we are cognizant that, as 
an interim solution, that end date is indeterminate at this 
point. We are making investments in VistA to make sure that it 
is resilient, that we maintain the level of performance that we 
have today. We are not stopping work on VistA. We realize it is 
going to be around for a long time.
    Mr. Self. The five that will remain, are they important? 
Are they minor?
    Mr. McCune. Those five, sir, will be replaced either in the 
short term or the long term. That list has been a little bit 
fluid, and so we are waiting for it to stabilize before we take 
action on that. If we do indeed stabilize on those five 
modules, we will start to work on modernization plans around 
those five.
    Mr. Self. Thank you, Mr. Chairman. I yield back.
    Mr. Rosendale. I would like to thank this panel for joining 
us today. This panel is excused from the witness table.
    We will bring the second round up.
    Thank you very much.
    Mr. Rosendale. I would now like to welcome the witnesses to 
our second panel to the witness table. I ask you to please 
stand and raise your right hands, Mr. Baker, Mr. Gfrerer, Mr. 
Levin.
    [Witnesses sworn.]
    Thank you so much.
    Let the record reflect that all witnesses have answered in 
the affirmative.
    Mr. Baker, you are now recognized for 5 minutes to offer 
your opening statement.

                    STATEMENT OF ROGER BAKER

    Mr. Baker. Thank you, Chairman Rosendale and Ranking Member 
Cherfilus-McCormick, for holding this hearing today.
    With over $50 billion at stake, misinformation regarding 
VistA has been rampant. Numerous parties continue to repeat 
this misinformation in an effort to convinces Congress that 
VistA is a problem in need of their solution.
    The primary success measurement for an electronic health 
record system at VA should be veteran health outcomes. That is, 
after all, the fundamental reason the VA and the EHR exist.
    The EHR program has effectively run a controlled experiment 
over the last 6 years, complete with a hypothesis, a control 
group, and metrics. This experiment has provided concrete proof 
that veterans achieve better medical outcomes when VA 
facilities use VistA than when they use the alternative.
    Because VistA excels in medical care, it sets a high 
comparative bar. VistA, and the work processes encoded in it, 
was designed, implemented, and honed by VA clinicians to do 
exactly what a clinician needs and exactly what a clinician 
wants. That is both its blessing and its curse.
    Independent surveys have shown that VistA is the most liked 
her by clinicians nationally, and that is among all EHRs.
    Lobbyists would have you believe that the IT difficulties 
of VistA are more important than the medical care advantages. 
VistA is a problem because of its age, its complexity, and the 
language it is written in. In fact, at least one lobbyist would 
have you believe VistA cannot be made better.
    That is provably false. During just my 4-year tenure, many 
improvements were introduced to VistA, including bed 
management, blood bank, pharmacy reengineering, registries, and 
numerous others, notably Cerner Labs.
    In fact, VistA can be difficult to modernization. The 
difficulties in modernizing VistA stem not from the software, 
but from three root causes that come directly from VA itself.
    First, every time VA has attempted to replace VistA, 
starting in the year 2000 under the HealteVet program and from 
2017 under EHRM, VA has prohibited further modernization of 
VistA. This has included eliminating promising technology such 
as VistA Exchange and Enterprise Health Management Platform 
(EHMP) even when they were already in beta test. The fact that 
you are told that VistA cannot be made better, when the primary 
barrier to modernizing VistA has been a lack of investment for 
16 of the last 24 years, is rich indeed.
    Second, years ago VHA made the decision that veterans 
receive better medical care if each VA is allowed to tailor its 
care to local needs, that veterans in Fort Harrison VA can have 
different medical needs from those in West Palm Beach. This 
local control of medical care is a fundamental part of the 
medical culture of VHA, and VistA is reflective of that 
culture. This is where the sound byte, ``This is not a single 
system, it is 130 systems'' comes from. Local customizations, 
including local development staff, were a celebrated part of 
VistA development for many years until the advent of the Gold 
Disk Program in 2011. These local customizations are what has 
made designing, programming, and testing changes to VistA more 
difficult because every change must be tested to work with each 
VistA instance.
    Third, Federal pay grades and procurement practices have 
eroded the base of skilled software developers needed to 
maintain a complex her product. Capping salaries at GS 14 
levels for the most skilled Federal IT staff has caused them to 
seek other employment. VA continuously awards contracts for 
complex VistA improvements to companies that cannot employ the 
necessary skills at the rates that were bid to win the work. 
They would rather tell VA: We can not find MUMPS programmers 
without adding at those rates to justify why they can not 
deliver.
    VA has repeatedly failed at efforts to replace, not to 
modernization VistA. Unless a decision is made that software 
standardization is more important than local control of 
healthcare at VA, attempts to replace VistA with a product that 
does not support that fundamental part of the VHA culture are 
doomed to failure.
    Mr. Chairman, misinformation regarding VistA is being 
promulgated to justify the now $50 billion need for the EHRM 
program. The EHRM program itself has provided the best proof 
that they are untrue. After 6 years veterans continue to 
achieve better healthcare outcomes in VA facilities that use 
VistA. That remains the single most important fact that you 
will hear today.
    I commend this committee for demanding the actual facts 
regarding VistA, its role in veteran healthcare, and its 
ability to be modernized. I look forward to working with you 
and answering your questions as you further search out these 
facts.
    Thank you.

    [The Prepared Statement Of Roger Baker Appears In The 
Appendix]

    Mr. Rosendale. Thank you, Mr. Baker.
    The written statement of Mr. Baker will be entered into the 
hearing record.
    Mr. Gfrerer, you are now recognized for 5 minutes to 
deliver your statement.

                   STATEMENT OF JAMES GFRERER

    Mr. Gfrerer. Thank you, Chairman Rosendale, Ranking Member 
Cherfilus-McCormick, and the subcommittee today, for the 
opportunity to appear.
    As a veteran I am a patient in VHA's health system. I am a 
beneficiary in the benefit system, and now preregistered in 
VA's burial benefits. As more than 28-year career Marine 
infantry officer, with four combat deployments, I fully 
empathize with all of our veteran men and women who endure both 
the visible and invisible wounds of military service.
    There is much misunderstanding around VA healthcare in 
general. VA healthcare is unlike commercial systems. VA is 
funded by government appropriation versus commercial health 
systems who operate on a business revenue model. In commercial 
healthcare, each patient is eligible for all services, where in 
VA, eligibility is based on complex service-connected 
conditions. VA healthcare is more specialized and expansive 
than commercial systems comprising unique capabilities, such as 
prosthetics, long-term care, and dental among others. These are 
substantial differences even as compared to Department of 
Defense healthcare and the first set of challenges for any 
commercial her to be successfully implemented into VHA.
    The bottom line is that Federal law, regulation, and policy 
have created this unique health system, and the VistA her is 
representative of those complex and unique business rules. It 
may come as no surprise that when a commercial her programmed 
for different financial frameworks, with significantly 
different eligibility rules, and not addressing unique VA 
clinical services, that there are problems and problems that 
can not be overcome by change management. Without substantial 
customization, no commercial her could address the business 
rules that law, regulation, and policy mandate for veteran 
healthcare. If you did not have a business system configured 
like VistA, you would have to create or heavily customize a 
system to perform just like it.
    In the remainder of this hearing, we will get into greater 
detail about VistA, its modernization efforts and some 
additional facts and misconceptions, but allow me to offer some 
highlights:
    First, VistA is more than an her. It is what professionals 
term an Enterprise Resource Planning or ERP system, which has 
grown over the years to encompass many administrative, 
financial, and other modules, a number of these which will live 
on.
    Second, it is not--I repeat not--a, quote, IT system but 
rather a business and mission system. Why does this matter? 
First, because the business, in this case VHA, must take prime 
ownership to include the life cycle management, the capital 
investment, and change management, with OIT playing a 
supporting technical role.
    Third, some would have you believe that VistA has not been 
modernized, but that assertion is predicated on the fallacy 
that modernization can only be achieved by replacement. Tech 
modernization as defined by Gartner, Forrester, and others can 
be--they say can be achieved in a myriad of other ways, 
rehosting, moving to the cloud; refactoring, optimizing the 
existing code; and encapsulating, exposing to APIS, all of 
which were done to VistA during my tenure.
    Also, let me offer in many respects veteran health care 
business and technology discussions remain mired in the 2017 
timeframe. It was in this timeframe that the pursuit of a fully 
longitudinal health record was revalidated with an assumption 
that it must be on the same platform in order for this to be 
achieved. I will tell you in 2023, with the maturity and 
adoption of health information exchanges and health standards, 
that is no longer the case.
    In an era of increasing technical debt and mounting 
technology modernization costs, the Congress must determine 
where the greatest need is for precious taxpayer dollars. 
Presently there are roughly 300,000 Active Duty members who 
matriculate from military service to VA every year. Last year 
on the community care side, VA saw 6 million referrals for 36 
million episodes of care. To compare 300,000 one-time transfer 
to those staggering numbers, there is no doubt that the latter 
is the substantially larger problem set and needs to be 
addressed.
    Finally, in an era where technology plays an increasingly 
mainstream and critical role in healthcare delivery, VA must 
begin to operate more efficiently and effectively, as do its 
commercial and non-profit healthcare system partners, who are 
well on their way in this regard. These systems understand that 
technology and information technology is the success path and, 
reciprocally, health systems cannot hire their way out of the 
problem much as VHA attempts to do every year.
    Mr. Chairman, thank you and the subcommittee for your time 
today, and I look forward to your questions.

    [The Prepared Statement Of James Gfrerer Appears In The 
Appendix]

    Mr. Rosendale. Thank you, Mr. Gfrerer.
    The written statement of Mr. Gfrerer will be entered into 
the hearing record.
    Mr. Levin, you are now recognized for 5 minutes to deliver 
your opening statement.

                    STATEMENT OF PETER LEVIN

    Mr. Levin. Thank you for the privilege of testifying before 
you today regarding the Electronic Health Record modernization 
at the Department of Veterans Affairs.
    I am deeply grateful to you, Chairman Rosendale, Ranking 
Member Cherfilus-McCormick, and members of this subcommittee, 
for the opportunity to share with you my perspective on one of 
the largest civilian information technology projects in 
history.
    Our commitment to our Nation's veterans transcends party 
lines and political idealogy. In an era of especially deep 
ideology divide and social tension, I applaud your leadership, 
Mr. Chairman, in soliciting the best ideas and constructive, 
objective, fact-based perspectives from across the spectrum.
    During my time in public service and under the leadership 
of Assistant Secretary Baker, with whom I am deleted to appear 
today, this afternoon, I have the honor of working on several 
medical information technology systems that are still in use 
today. Especially relevant to this testimony are the Joint 
Longitudinal Viewer, JLV, originally known as Janus, and the 
Blue Button personal health record that was launched during a 
democratic administration and was warmly embraced by the most 
recent Republican one as a fundamental component of any effort 
to empower patients in their healthcare decisions. JLV enables 
hundreds of thousands of clinicians to see health records 
across platforms every day.
    In this context, health aid interoperability, I 
respectfully offer my observations.
    There are three issues before the government regarding 
VistA and Veterans Affairs: One, that billions of dollars 
already spent on the commercial implementation will not scale 
to enterprise wide clinical care services on the current path, 
budget, or timeline. Two, that VA can and should sustain the 
data interfaces and connection frameworks already built to send 
and receive data from Military Health System (MHS) GENESIS. 
Three, and most important of all, that VA consolidate its 
current instances of VistA onto a VA-centered clinical 
workflow, and augment the VistA model to receive data from 
third-party providers.
    As you will hear from my colleagues and other witnesses, 
the differences in VA healthcare between points of care is 
simply not that large. It is not tens of billions of dollars 
large. Veterans receive terrific healthcare, but their care is 
delivered with different processes depending upon which 
hospital or clinic they go to. That is the primary problem.
    In my opinion, the department should not announce its 
intention to change the contract unless and until it has a 
backup plan in place. That plan cannot be to, quote, revert 
back to VistA in its current form, or anything that concedes to 
VA's continued digital isolation and process insularity. 
Switching back to VistA and walking away does not fix the root 
problem. How do we address this issue?
    First and foremost, cloud technologies are now stable and 
mature enough to enable consolidation onto an authentically 
single platform. Additionally, there have been substantial 
improvements to the code base that are now available to VA from 
the commercial sector. It would be straightforward to 
reinstantiate Open Source Health Record Alliance (OSEHRA) with 
a powerful charter and legislative mandate.
    Moreover, commercially available data management 
infrastructure has made substantial progress in the last 5 
years. There are no technology impediments here. I would like 
to repeat that. There are no technology impediments here. This 
is simply a matter of political will, imaginative leadership, 
and execution accountability.
    Indeed, these policy changes, consolidation onto an 
enterprise clinical workflow and adoption of proven platform 
and data management services would accelerate health record 
modernization at a fraction of the cost now earmarked for EHRM.
    The transcendent goal of our work has to be the 
continuously better care of healthcare of veterans. Unless and 
until VA resolves its internal tension around consolidated care 
pathways and comprehensive data management, no amount of 
technology will automate its operations and no amount of money 
will solve its policy problems.
    Thank you for my opportunity.

    [The Prepared Statement Of Peter Levin Appears In The 
Appendix]

    Mr. Rosendale. Thank you very much, Mr. Levin.
    The written statement of Mr. Levin will be entered into the 
hearing record.
    We are now going to proceed to questioning and, I will 
start with the first 5 minutes' worth.
    Mr. Baker, I have heard it said that VistA was developed as 
a healthcare delivery software and that the Cerner Oracle 
system was developed as a billing software. What is the basic 
differences of that? What problems does it present?
    Mr. Baker. I think it goes back to the environments they 
grew up in. On the commercial side, billing is where her is 
started, to be able to accurately bill and get paid. On the VA 
side, they did not have to worry about billing out for most of 
the tenure of VistA.
    VistA focused on keeping a record that was completely 
oriented around how good is the medical care we are providing 
this veteran, whereas commercially EHRs tend to be excellent on 
the financial side and have added more recently the medical 
record side of things, at least to the extent that VistA has.
    The difference there is clearly on the focus. You know, as 
I mentioned in mine, there is a culture difference between the 
VA system and the commercial systems related to the preeminence 
of medical care. Congress provides an appropriation that 
provides most of the money that VHA needs. They do not have to 
worry as much about am I billing things out, am I going to be 
able to pay people as a commercial her does. They both have 
very valid reasons for why they grew up. What we are trying to 
do now with EHRM is to take a financial system and lay it over 
top of a medical records system.
    Mr. Rosendale. I think that very clearly describes exactly 
the frustrations that I have had since I was exposed to this, 
and that is that the benefit, the healthcare benefit is not 
being provided properly. I think that this has been expressed 
by the docs and physicians at the facility at Mann-Grandstaff.
    Mr. Baker or Mr. Levin, there have been previous efforts to 
modernize VistA as well as replace it. First of all, in the 
2000's, there was replacement effort. I understand it cost 
about $600 million.
    What was this project's genesis, how did it perform, and 
what happened to it?
    Mr. Baker. Well, since I was the person who terminated the 
HealtheVet program, I will speak to that one.
    In roughly the 2000 timeframe, VA decided that it could 
replace VistA by developing a new system called HealtheVet, and 
they launched down that path. A good friend of mine was the 
deputy Chief Information Officer (CIO) at that time, and his 
responsibility was to terminate all spending on VistA so that 
it could be focused on HealtheVet.
    There were a few good pieces of software that came out of 
HealtheVet. The My HealtheVet website that you see now came out 
of that, as well as a few other things that we were able to add 
to VistA. By and large, that $600 million, by the time I got 
there, it was obvious that was a wasted program, and so we 
terminated it and focused on the VistA improvements that I 
mentioned in my testimony.
    Mr. Rosendale. Very good.
    Then, Mr. Baker, next came the Integrated Electronic Health 
Record (iEHR) between 2011 and 2014. VA and DOD attempted to 
build a new joint EHR. I understand that that one cost about $1 
billion.
    How was that different, and why was it abandoned?
    Mr. Baker. I was very key and core to iEHR, and I would 
describe iEHR this way. It was a battle between VA and DOD. VA 
had a requirement that the only her they could use was VistA. 
DOD had a requirement that they could use any her but VistA. 
And we spent a lot of time in that battle.
    Now, let me--since Dr. Levin sees that time much more 
rationally than I do, given how deeply I was involved, let me 
just ask him if he has any further comments there.
    Mr. Levin. Thank you, Mr. Baker.
    I do have a couple of comments. I think that there is a 
gigantic difference that we ought to acknowledge that we have 
tried to replace VistA a couple of times, but our modernization 
efforts, looking at this holistically and trying to module by 
module, object by object replace it, upgrade it, modernize it, 
transfer it to the cloud, this has never actually received full 
attention and support from the institution.
    I think one of the conversations that we are having today--
and iEHR is a perfect example of that--is to distinguish and 
differentiate what is a replacement and what is a 
modernization.
    The single biggest problem with VistA, at least in my 
opinion, has nothing to do with MUMPS or the availability of 
language competencies of any individual programmer. As 
Secretary Baker has pointed out many times, if you speak 11 
computer languages, you can learn the 12th one without too much 
extra effort. I think that that is a little bit of a red 
herring to throw out the particular language and expression of 
the system.
    The number one handicap of VistA today is its lack of 
modularity. That is a word that Secretary Baker led when he was 
in office and I participated in, and I think that VistA itself 
or any her would benefit tremendously from attention on its 
modularity, its extensibility, its scaleability, its 
testability. These are things that are fixable inside the 
incumbent program today.
    Mr. Rosendale. Thank you very much.
    Representative Cherfilus-McCormick.
    Mrs. Cherfilus-McCormick. Thank you, Mr. Chair.
    Mr. Levin, in your written testimony you stated that the 
department should not announce its intention to change the 
contract unless and until it has a backup plan in place.
    Furthermore, you stated that that plan cannot be reverted 
back to VistA in its current form or anything that concedes to 
VA's continued digital isolation and process insularity. I just 
want to State I cannot agree with you more.
    The EHRM program is one particular case study that is part 
of a larger issue surrounding the VA's inability to manage and 
contract for large IT systems. Whether it is EHRM, supply 
change, or HR modernization, the lack of progress and success 
is not contained through a single program. Without real program 
and acquisition management at VA, this program will continue to 
not be successful.
    Mr. Levin, would you be supportive of holistic changes with 
how VA manages and is held accountable for these large IT 
modernization efforts?
    Mr. Levin. Thank you very much for your question, 
Congresswoman.
    The short answer is yes, I could not agree with you more. I 
would like to introduce you to some members of my family to 
find somebody to agree with.
    The things that we could do structurally operationally 
inside the VA I think are very accessible to us. I can not 
explain why we do not do them on a regular basis, but written 
down plans, accountability metrics, performance metrics--and I 
do not just mean by individuals; I mean by the systems that we 
are putting in place. I think that we have confused--and we can 
have probably a larger societal discussion about this. We have 
confused the difference between best efforts and outcomes. I 
think that, at least in my world, the commercial world that I 
live in today, my customers do not very much care how hard I 
work. They very much care about whether I deliver to my 
promises on time.
    There is a huge difference, a vast difference between how 
the processes and the accountability metrics, the hiring that 
we have to go through at any government institution. Of course, 
VA is not unique in that way at all. We can do better. There 
are lots of things in the private sector, obviously, that I 
think we would agree are sometimes inequitable or sometimes 
ineffective. The pendulum has swung too far, in my opinion. In 
my personal opinion, the pendulum has swung too far, and we 
should be able to structure programs inside of government, 
inside of VA that hold the leadership accountable, that have 
sensible budgets associated with them, and that have 
performance-measured outcomes. I do not think that we are doing 
that sufficiently well right now.
    Mrs. Cherfilus-McCormick. Thank you.
    Given that EHRM was allowed to be awarded as a sole source 
contract, do you think that the VA's requirement requires more 
checks and balances internally before awarding future contracts 
for IT modernization given the results that we have seen thus 
far?
    Mr. Levin. Again, thank you for your question.
    I believe that my copanelists will have divergent views on 
this.
    I was relatively close to the process. As a private sector 
observer, I was relatively close to the process. I do not think 
that it was unfair. There was political exigencies. There were 
operational exigencies at the time. I think that if we could 
wind back history, it was not so much that we made a decision. 
I was a very loud advocate of not going the direction that we 
went, but I had the privilege of speaking to a senior most 
leader at the time who was part of the decisionmaking and who 
reported to me: okay. The decision is made. Now what?
    I said to that individual: Well, now it is just an 
engineering problem. Anybody can do that. Right? I still 
believe that. I really do.
    I think that--if I can recommend attention of this 
committee, the focus of this committee would not be so much on 
how do we get to the DNF and why did we follow the DOD. Those 
are important questions. Those are policy questions. It is what 
all happened afterwards, where many of us who were loudly 
advocating the misdirection, the misapplication, the lack of 
framework, the lack of accountability, the personalities that 
were involved, and there was almost no accountability 
whatsoever, I think that that is where I would direct this 
committee's attention, and I think that that is what we could 
be doing much better at VA.
    Mrs. Cherfilus-McCormick. Thank you so much for your 
testimony.
    Mr. Chair, I yield back.
    Mr. Rosendale. Representative Self.
    Mr. Self. Thank you, Mr. Chairman.
    I want to go back to the languages, though. That seems to 
be a major point for the VA.
    As we have said, other old languages are used. Is there any 
other language that could be used to modernize VistA?
    Mr. Baker. As modules are changed in VistA, frequently they 
are written in other languages. In my written testimony, I talk 
about VA's attempt to implement Cerner Labs, which is written 
in a completely different language.
    The points at which MUMPS is critical is as you understand 
the complexities and the architecture of VistA and making 
changes to that as you add the new modules that are going to 
interface to it. That MUMPS expertise is necessary. Even in 
saying MUMPS, what you are really saying is expertise about 
VistA as a large-scale application. It is, by definition, by 
what it tries to do, a very complex piece of software because 
it is also a very accurate and very extensive piece of medical 
software. We want it to be as good as it is because we get our 
medical care from that.
    The short answer is yes. Frequently the new pieces are 
written in different languages, but you have to have some level 
of MUMPS expertise to be able to interface to what remains 
there from the VistA system.
    Mr. Self. If I understood you correctly, surely there are 
people, other than your more mature MUMPS coders, that 
understand the VistA architecture?
    Mr. Baker. It is a--it is not something that you want to 
give a neophyte access to. It is like having them drive, you 
know, your Maserati, having your 16-year-old son drive your 
Maserati, not a good idea.
    You would like them to get some training on other things. 
Again, a lot of times the problems that VA runs into from a 
procurement standpoint is that they do not--they award 
contracts when even those senior level people are necessary to 
guide the more junior people, they are just too expensive for 
that contract. You end up with a solution that does not work.
    Mr. Self. I understand that clinicians like VistA; IT does 
not. What are the IT reasons to cancel VistA? Then we will get 
to the clinicians.
    Mr. Gfrerer. Congressman, I would probably reject that 
characterization. I mean, IT is--certainly what is not to like 
is that when you are working with riddled systems, you know, 
that are not, you know, refactored and hosted and, you know, 
the most--the best environments, right. It is a work in 
progress on VistA. Ultimately, when you are back in 2017--and I 
was admittedly not with the agency--it was an agency decision 
to go to VistA. It was not a vote of no confidence within OIT.
    I think the other thing too is, to get back to your 
language point, at some point, you know, the increasing trend 
you see in all verticals, in healthcare, and others, is managed 
services where companies that do this work well can take on 
bits and pieces, those modules that we are talking about. You 
are going to see more of that, right, companies that do 
pharmacy real well, companies that do scheduling real well. You 
know, the VA should not be in that business. Right? They should 
look to who are the vendors that are best of sweet with regard 
to these capabilities and incorporate them. At that point you 
have shifted the risk to the vendor, to the commercial 
provider, and you do not care if it is written in MUMPS or C 
Plus or COBOL, or whatever language it is in.
    Mr. Baker. If I could add, sir, the interesting point the 
chair was making earlier about same number of staff, the IT 
people that work on VistA love the product. They love the 
product. The affinity--the people that I have heard from inside 
VA since they found out I was going to testify telling me, you 
know, please do this, has been extensive.
    There is this large group of people, several thousand, that 
know this product well and want to see it succeed. They are 
part of what we looked to engage with the open source community 
at the time we were doing that, and they would like to continue 
to participate in moving this product forward. They have a lot 
of expertise.
    Mr. Self. Interesting.
    That brings me to a simple question. What advantages to the 
veteran does Cerner bring?
    Mr. Baker. I do not think I am going to be somebody who is 
going to give you any of those. I am not aware of things that, 
for the veteran today, are better because of the Cerner 
implementation in the five facilities than they were before 
VistA was there.
    I will ask if my partners want to say anything.
    Mr. Gfrerer. Yes. Maybe as the lone veteran, I will hazard 
a guess here and say that, ultimately, it should be about the 
clinician deciding that the care outcome is going to be better 
for the veteran. Largely, as the veteran--and I saw this in 
some of the earliest communication about how great the new 
Electronic Medial Record (EMR) was. Again, as a veteran 
patient, it does not matter to me. I should not know whether it 
is VistA or, you know, whatever the EMR is that the care is 
being provided for.
    At the point that I care about what the system is that is 
being provided, I am probably being negatively impacting.
    Mr. Self. I am speaking to former secretaries. In your view 
or in the view of the current VA, I just--I am not--I am 
starting to not understand any advantages of Cerner.
    Mr. Baker. I am not aware of any----
    Mr. Self. I am not talking--I understand you are a veteran, 
I am a veteran, and it should be transparent to us. But at your 
level, at the decisionmaking, the policymaking level, what are 
the advantages that Cerner will bring?
    Mr. Baker. The theoretical advantages--I do not believe 
they have been realized yet--is that Cerner would be easier to 
maintain and easier to move forward than VistA in the out years 
of the program. That would be the theoretical advantage, but 
there is no hard evidence that that theoretical advantage will 
actually be true. The cost, in my opinion, is going to be 
substantially more than it would be, from a VistA perspective, 
for the near and the midterm.
    Mr. Self. Okay. You have just gone back to my first 
question: What is the IT advantage of Cerner? There does not 
seem to be one.
    Mr. Chairman, I yield back.
    Thank you.
    Mr. Rosendale. We will come back around for another round 
so you can have some more.
    Now we are starting to get to the crux of where I wanted to 
go for this next round of questioning, and that is the 
investment as opposed to the theoretical promise of benefit in 
the future; okay.
    Mr. Baker and Mr. Levin, let us turn now to the 
modernization efforts. What was the purpose of the Gold Disk 
project? How much did it cost? What did it accomplish?
    Mr. Levin. Secretary Baker, I do not remember how much it 
cost, but I do remember it not being very expensive. Is that 
true?
    Mr. Baker. My best understanding is it was single digit 
millions, probably in the 5--I would want to ask VA for the 
specific numbers, but it was not an expensive program.
    Mr. Levin. I would have hazard--I am glad that Secretary 
Baker went first. I would have guessed it was much smaller than 
that.
    The outcome of the Gold Disk Program, which I want to 
immediately confess, was not perfectly successful. It was a 
great first start. There was a tremendous amount of resistance 
to consolidating that because, of course, that was the 
premonition of consolidating workflows and understanding the 
differences and highlighting the variations in the clinical 
care pathways at the various installations.
    I do not want to make it sound like the Gold Disk was, in 
fact, what its name implies, the nomenclature amplifies a 
little bit the success. That not withstanding, probably the 
most important outcome--I am looking at Secretary Baker when I 
say this--was that we consolidated FileMan.
    The most challenging, the most difficult, arguably the--
not--if we use an anatomical analogy, it was not the heart and 
brains, but it was one of the vital organs. We went after it 
first because it was difficult, because we wanted to 
demonstrate that, in fact, you could start modularizing the 
code, you could start creating test harnesses around the code, 
that you can do what we are discussing here today. This goes 
back 12 years ago, sir.
    Mr. Rosendale. Then for several million dollars, you were 
able to accomplish this?
    Mr. Levin. Yes.
    Mr. Rosendale. Okay. Mr. Baker, Mr. Levin, why were you not 
able to eliminate all of the differences between the VistA 
instances?
    Mr. Baker. Well, the most important thing was it was a 
joint program between VHA and OIT and we got to the point where 
there were differences that VHA was unwilling to, if you will, 
exert their internal political capital to get their 
organization to agree to. When we achieved--we had gotten to 
about 95 percent commonality across the platform when I left. 
We felt that was a good start, and we felt it would continue 
on.
    I want to commend Mr. McCune. My understanding is that it 
has continued on and that they are now achieving, if not a 
single Gold Disk, but very close to that with VistA at this 
point. His efforts on that part have been great from our 
perspective.
    Mr. Rosendale. Very good.
    Mr. Baker, Mr. Levin, VistA Evolution began in 2014, and it 
was canceled in 2017 when Cerner started. It was estimated to 
cost about $5.3 billion over a decade.
    What were its goals and what was accomplished with that 
program?
    Mr. Baker. I do not know that I can speak to the goals. I 
think VA's goals were bigger. It was developed after I left, 
and so I think the goals were bigger.
    I actually was at one of the contractors at that point that 
was developing that. What was developed was something called 
VistA Exchange, the ability for a clinician--the data provided 
to a clinician to come from a wide variety of data sources, 
multiple EHRs, multiple data sources, and present a single 
computable view to the clinician and a web-based version of the 
CPRS interface, so a modernization of that into what is called 
a widgets-based interface.
    Both of those were in beta tests, and other things were in 
beta tests at the Hampton Roads facility when that project was 
canceled.
    Mr. Rosendale. Okay. This question is going to be for all 
three witnesses.
    Regardless of the EHR system chosen, is the VA ever going 
to be able to accomplish a monumental rip-and-replace project? 
Should they even be trying to do that?
    We can start over to my right, Mr. Levin.
    Mr. Levin. Congressman, with respect, I believe that the 
answer to your question is no. It is not so much even a 
question of the competency and skills. It is certainly not a 
question, not a question of the commitment of the people who 
work at VA. It is absolutely a question of the structure, the 
accountability metrics, the frameworks in which they have to 
operate today.
    Under its current framework and structure, the answer is 
no.
    Mr. Gfrerer. Over the past decade, commercial entities have 
gone away from big ERP, Enterprise Resource Planning, rip-and-
replace. It is just fraught with too much peril and failure. 
The incremental modernization path is generally the one that 
you see followed.
    I would like to layer in too at some point that, you know, 
any organization, whether it be the VA or a major healthcare 
system, really only undergoes these sorts of, you know, 
changes, you know, once every 30 to 40 years. I mean, it is 
almost like big iron in the sense of they are there--you know, 
look at VistA, it was there for 40 years. It is still there.
    The knowledge base around putting in a new system in 
whatever manner, it is fairly limited, fairly exclusive. It is 
not astronaut exclusive, but there is only--when you look 
across the executive base in the U.S., there is only, you know, 
dozens of people that have really done this. I can tell you 
that at VA during my time and others' times there are none of 
those people. Right.
    The requirement to leverage external advisors and to visit 
and--you know, I can talk more about this later--I do not think 
that has been anywhere near sufficiently leveraged. I know when 
I compare during my time the level of governance that was 
exerted on the program compared to what I have talked to some 
of my commercial counterparts, it is not even a shadow of what 
they did in term of these programmatic deployments.
    Mr. Rosendale. Thank you.
    Mr. Baker.
    Mr. Baker. No is the simple answer. The government in 
general is not good at large IT programs. There was a point in 
time in about 2010 when the National Defense Authorization Act 
included the language that 16, 1-6, percent of large Federal IT 
programs succeed.
    The EHRM is the largest Federal IT program. It is not being 
done by DOD. VA does not have the same program management 
capabilities that DOD has. Especially looking at the evidence 
of the growth in the budget of the program, it is a very good 
indicator that, at this point in time, VA has no chance of 
actually succeeding on this program.
    Mr. Rosendale. Thank you. Thank you very much.
    Ms. Cherfilus-McCormick.
    Mrs. Cherfilus-McCormick. Thank you, Mr. Chair.
    Mr. Gfrerer, what role did you play in Cerner's EHR 
modernization at the VA?
    Mr. Gfrerer. I feel like I am back in my confirmation 
hearing.
    Certainly the Office of Primary Responsibility for EHRM and 
VA very deliberately was segmented into a specific office, 
during my time the Office of Electronic Health Record 
Modernization. Then it was supported by two business units that 
were the existing EMR owners; VHA as the business owner and OIT 
as the technical owner and operator of the system.
    Both VHA and OIT supported the Office of Electronic Health 
Record Modernization and their programmatic oversight and 
scheduling of those waived deployments.
    Mrs. Cherfilus-McCormick. Do you think in retrospect that 
was a bad decision to have OIT on the outside of the program 
with CIO having such a limited role?
    Mr. Gfrerer. When you look at innovation and putting in new 
applications and new systems, often you have to get past the 
current, you know--the existing system owners, both the 
business and the IT. I do not think that that was necessarily a 
bad decision. You know, each organization does it a little 
differently.
    I had a conversation with a commercial CIO, as just kind of 
a refresher for this panel, the other day. He elected in, like, 
his own cylinder of excellence to have a separate organization. 
There is pluses and minuses to both approaches, and that is the 
one the VA selected.
    Mrs. Cherfilus-McCormick. I recognize you entered VA 
service a few months after the Oracle Cerner contract was 
signed. From your perspective now, do you think the strategy to 
do a sole source contract with Oracle Cerner was a good 
decision?
    Mr. Gfrerer. Well, as I said in my opening remarks, in that 
timeframe, the premise when you looked at where technology was 
or, more importantly, where it was not in terms of Health 
Information Exchange, again, you could make a case that if you 
are going to derive the benefit of having that longitudinal 
record, you know, all the way from the time someone enters 
military service until the time they are buried in a VA 
cemetery, that it needed to be on a single platform.
    That is why I said a lot has changed in the past 6 years, 
certainly the development of the Health Information Exchanges, 
and then when you layer in this whole notion that you have this 
smoldering ember of a problem over here with 300,000 people 
changing hands every year and over here you have 6 million 
community care, you know, referrals,.
    Thirty-six million episodes of care, that is only going to 
grow larger. We need to kind of make a rational decision about, 
you know, where those precious investment--those tax dollars 
are paid.
    By the way, where you are having that progress on the Joint 
Health Information Exchange, that can also accrue benefits back 
into the DOD VA system and that data exchange.
    Mrs. Cherfilus-McCormick. Do you think the VA did 
appropriate planning, research, and standard program management 
related to this contract and program?
    Mr. Gfrerer. As I mentioned just a few minutes ago, when 
you compare and contrast what the VA has done and likely 
continues to do and compare it to any one of the number of 
large commercial systems that have been on a similar electronic 
health record journey, you would be concerned about the 
disparity between the governance and the efforts that these 
commercial systems have exhibited compared to VA.
    It is what Secretary Baker said too is, by and large, not 
to give an excuse, government agencies do an exceptionally poor 
job at managing. I would call them--I said it before--business 
transformations. They are not IT programs. It is not an 
enterprise IT. It is not a security program. We are not putting 
in a new service management tool. This is a business admission 
healthcare system program.
    Mrs. Cherfilus-McCormick. My final question for you is do 
you think it was a wise decision to deploy Cerner EHR in the 
middle of the pandemic, a facility that was not yet ready, with 
technology that was not ready, just a few weeks before a 
Presidential election?
    Mr. Gfrerer. Well, I know that the original date, having 
been in those meetings, was October 2019. The decision at that 
point on go, no go was really predicated around--what the 
chairman and others have talked about was those 73 different 
applications or interfaces that Millennium was going to have to 
draw services from. Those were just not ready. That was the 
initial decision to push off.
    Then at the point that it went into the following year, it 
really was more in VHA and to the facility, to Spokane's 
decision as to when they would be ready to go live. I think 
that decision was really pushed down to the lowest level 
possible, certainly to the medical center director at Spokane.
    Mrs. Cherfilus-McCormick. Thank you for your testimony.
    Mr. Chairman, I yield back.
    Mr. Rosendale. Representative Self.
    Mr. Self. Thank you, Mr. Chairman.
    Given you all's history with VistA, order of magnitude, 
what would it take to rewrite VistA in dollars? Order of 
magnitude.
    Mr. Baker. I would hope I am on the high side with $10 
billion.
    Mr. Self. Thank you.
    Mr. Chairman, I yield back.
    Mr. Rosendale. Thank you, Representative Self.
    I am going to go into these instances. I have got a batch 
of questions about that. This question is for anyone who wants 
to answer.
    Why do the different VistA instances exist? How are they 
different? How much of--what I am trying to figure out is how 
much of each instance is identical, okay, that they all match 
up, and what are the examples of the differences of the 
instances?
    Mr. Gfrerer. I will start off, Mr. Chairman. I am sure 
Secretary Baker and Mr. Levin will jump in.
    As you heard earlier, the rough average is about 95 percent 
commonality between the 133 instances. I would say, again, it 
is a business admission decision as to that variability; right.
    The first thing you have to realize is, you know, not every 
VA facility is the same. There are a range of clinical 
capabilities all the way from--let me just use the Northwest 
where it is currently being deployed, all the way from Puget 
Sound, which is a level 1a facility with all the top levels of 
care, Intensive Care Unit (ICU), Operating Rooms (ORs), things 
like that. Then you compare it to Spokane which is down at 
level 3. You can imagine that based on those different levels 
of facility care and those clinical capabilities, there would 
have to be some variability or differences in the instances.
    I will pass it off from there.
    Mr. Rosendale. Basically it is talking about the, so I am 
understanding, types of procedures that are offered in each of 
these facilities, some being much more complex and some being 
just a standard primary care?
    Mr. Gfrerer. I mean, literally different clinical 
capabilities. Again, you know, Puget Sound has an OR and an 
Emergency Room (ER) and I do not believe----
    Mr. Rosendale. Outside of that, let us just say outside of 
that, if we have similar--we will call it similar facilities 
offering similar procedures, okay, what are the differences? 
Are we still dealing with 95 percent commonality, or is it 
greater?
    Mr. Baker. I think today you are dealing with more. At the 
time that I left off, there were a few packages where there 
were alternative versions of the package that one facility 
might choose to implement and another facility would choose the 
alternative. Again, this comes from the day and time when the 
way packages got developed is one facility would build them, 
and another facility would build an alternative to that, and 
various hospitals would choose which of those they liked.
    By the time we started the Gold Disk Program, there was a 
plethora of what is called class 3 software.
    Mr. Rosendale. Were they all still communicating?
    Mr. Baker. Communicating to the sense that you can see the 
record from any VistA system in any other VistA system. I do 
not know whether that data is computable in all circumstances 
when it goes from one VistA system to another VistA system. It 
depends on your definition of, quote, communicating.
    There was a plethora of this class 3 software, and what we 
did was basically asked VHA to settle on one version of each of 
those packages and then make that the standard. They got down 
to about 5 percent where they said, We need to offer both of 
these or all three of these because we can not tell the 
facilities to just go with one.
    The variation was not, if you will, things all over the 
map. It was, oh, you have chosen the 1a option instead of the 
1b option for your facility, but we support both the 1a and the 
1b option across the enterprise.
    Mr. Rosendale. Okay. They were supporting them.
    To make sure that I am clear, so when I say communicating, 
so you could view it--regardless of which facility you were in, 
you could view it?
    Mr. Baker. Yes.
    Mr. Rosendale. Could you enter data--if a veteran came from 
one facility to the other, could you still enter data in if you 
were at a different facility?
    Mr. Baker. My understanding is that the veteran would have 
a new record, a new record at the new VA.
    For example, like I live in Florida. Lots of people will 
come down to live in Florida for 6 months and a day for tax 
reasons and, you know--but they will have a record in their 
Florida hospital, as well as their Ann Arbor hospital, for 
example, and you can view. They will have separate records 
between the two hospitals. That is my understanding, subject to 
correction by VA.
    Mr. Gfrerer. No, that is correct.
    I can use myself as example. While on Active Duty, I had 
a--had to be seen for a minor accident in Minneapolis. I had a 
record there, and I have a record in Department of Veterans 
Affairs Medical Center (VAMC) D.C., right, and the two are just 
related.
    Mr. Rosendale. Again, can either facility make amendments 
on either record?
    Mr. Baker. I do not----
    Mr. Rosendale. Or can you only make amendments in your 
facility for that record?
    Mr. Baker. I believe that you would have to specifically go 
to the other record and sign in as a user on the other VistA 
system to make actual changes in the other record is, I 
believe, the answer.
    Mr. Rosendale. Okay. Let me ask you this. At least does it 
come up as, like, a file, as an additional file so that you can 
see it as part of the patient chart?
    Mr. Baker. Yes. My understanding is that you get an alert 
as a doctor that says there are other records, other VistA 
records available on this patient. Would you like to view them?
    Again, I want to leave this subject to correction by VA. 
They know the system in depth much better than I do.
    Mr. Rosendale. Okay. Thank you. Thank you very much.
    I am out of time myself. I am going to turn this over to 
Representative Cherfilus-McCormick.
    Mrs. Cherfilus-McCormick. Thank you, Mr. Chair.
    This question is for Mr. Baker.
    You participated in VA's iEHR program, which eventually 
collapsed and the transition to VistA evolution, the program 
which has stopped, before the current Oracle Cerner EHRM, a 
very conservative estimate from the GAO using VA data show that 
efforts costing at least $1.1 billion.
    What makes you think going back to VistA is a good idea 
right now?
    Mr. Baker. I base my assessment on--of why you go back to 
VistA on two things that I talked about.
    One is just based on the track record at this point of the 
program. For someone like me, who has seen a lot of these 
programs, it is obvious this program has failed. The question 
is how much money we will spend before we fully recognize that.
    The second thing is one of the main metrics I use is are 
veterans getting better healthcare. If today veterans were 
getting better healthcare in facilities where the commercial 
product was being used, my view would be entirely different. 
You know, compared to the amount of money we spend on medical 
care at VHA, $50 billion is a relatively small amount. If you 
are not going to produce better healthcare, then why spend that 
money, because it is all about better healthcare for veterans.
    That is my reason why I believe we have got to go back to 
VistA and then a path forward. I do not think VistA is the end 
solution. I want to make that very clear. We are at this point 
now because we have failed on the EHRM, and we need a solution 
for the midterm while we figure out how to get to the long 
term.
    Mrs. Cherfilus-McCormick. Thank you.
    Do you have ideas for how VA could descope the current 
contract with Oracle Cerner and use it for the bear minimum 
services and then use other IT service providers and companies 
to fill in the gaps so as to diversify risk?
    Mr. Baker. I do. It would take us much longer than we have 
in testimony. I think about--sadly, I think about this a lot. 
Something about VA gets into your blood. I can speak to all 
three people here. We remain dedicated public servants even 
though we are not getting paid by the government at this point.
    Yes, we have talked about this. We have lots of thoughts on 
it. We would be happy to share those with you.
    Mrs. Cherfilus-McCormick. That would be wonderful.
    Given what you know about the IT industry, would you have 
done a sole source, no-bid contract like VA under the Trump 
administration did? Would you have mandated the use of an 
integrator?
    Mr. Baker. I am not a fan of integrators on large programs 
like this, just from all of the--from everything I have seen 
from that standpoint. My track record speaks to that.
    No, I did not like the sole source. I am loathed to 
criticize people that were in the arena at the time, but it was 
an odd way to go at things.
    Mrs. Cherfilus-McCormick. I read your written testimony and 
agree with you that the VA culture of allowing customization in 
its medical centers has contributed greatly to failed attempts 
at modernization.
    Is there any management structure in place at the VA where 
Under Secretary for Health and the CIO would be required to sit 
down and make decisions based upon priorities for deploying a 
modern EHR?
    Mr. Baker. I will just start by saying it depends on what 
administration you are talking about. Under Secretary Shinseki, 
absolutely, absolutely. I know it varied in other 
organizations.
    Mr. Gfrerer. I would add to that that I know, toward the 
end of my tenure, we attempted to re-implement the Office of 
Military and Veterans Affairs (OMVA) 130 Required Investment 
Review Board Process managing information as a strategic 
resource, and I can tell you that was met with not a lot of 
enthusiasm. That is--there is a reason that OMVA circular 
exists, because you have--and we saw it within our unfunded 
requirements process every year. You had 10X of requirements 
chasing 3X of IT spent dollars. If you did not deliberately, 
through an investment review board process, prioritize and be 
very clear eyed about where you were going to spend it, it was 
going to sort itself out and probably not in a very clean way.
    Mrs. Cherfilus-McCormick. I am looking for an opportunity 
to produce legislation this Congress to provide the VA with 
tools to make consensus decisions that put the care of veterans 
and the well-being of frontline staff at the forefront of 
requirements development and not cater to fractions within the 
VA. I prefer one software product over another. I do not care 
what system the VA uses. I want veterans and employees to be 
delivered what they are promised and what taxpayers are paying 
for.
    Thank you, Mr. Chair. I yield back.
    Mr. Rosendale. Thank you very much, Representative.
    Okay. Mr. Levin, I am going to ask a couple more questions 
before we let you all get out of here.
    VA leaders have said many times that they want an EHR with 
one instance, to eliminate the variations from facility to 
facility. Yet I hear regularly, almost every hearing that we 
hold, that if you have been to one VA, you have been to one VA 
because they are so different.
    My question is, is that realistic? How can it be 
accomplished?
    Mr. Levin. Thank you, Chairman.
    It is realistic. It can be accomplished.
    Let us spend a couple of seconds on how we got to now. 
There are endogenous reasons, internal reasons for the 
differences between the different facilities, and these are 
good reasons. These are things that we need to protect and 
nourish.
    The VA has a culture of innovation. We spend a lot of money 
every year with new medicines, new therapies, new processes, 
new instruments.
    Secretary Gfrerer mentioned during his remarks that it was 
a business decision. I strongly agree with him, and I strongly 
agree that that was the right thing to do.
    The VA is a greenhouse of training and new ideas, and you 
actually want that. We do not want to force or force feed some 
kind of rigid, unchangeable structure, because they are also 
teaching hospitals. Something like a third of all practicing 
physicians in the United States did at least some part of their 
residency at the VA. You need the VA to be vibrant that way, to 
be intellectually challenging, to show the hardest cases, to 
get the best care in order to attract the talent that, as 
Secretary Baker said--and I will say it slightly differently--
for many of us, those who are veterans and those who are not, 
those who are in public service today and those who are not, VA 
is a transcendent moral cause. It is a moral cause of our 
country, and I think that we made a promise that we need to 
keep. One of the ways that we keep that promise is by making 
sure that we are at the vanguard, at the forefront of that kind 
of care.
    The differences between the institutions, we have talked a 
lot about, well, how many of the clinical care pathways are the 
same or how much of the code base is the same. I respectfully 
submit that that may not be the right metrics. It is an easy 
metric. We know how to measure lines of code. We know how to 
articulate the differences.
    The difference that I would propose is that we see how much 
does it costs. How much does it cost to sustain these various 
pathways and do they actually deliver superior care? The trade 
space, the decision that we have to make together as a country 
is how do we sustain and maintain and nourish the vibrancy of 
those environments but not charge the taxpayer tens of billions 
of dollars because we are trying to let every VA hospital do it 
by themselves? At the end of the day, this is an implementation 
framework question. This is a question of authority. This is a 
question of accountability, and it is not a technology problem. 
I think that we can achieve it quite easily.
    Mr. Rosendale. It seems like the VA is counting on the 
Oracle Cerner EHR to force some kind of a standardization on 
the Veterans Health Administration without doing any of the 
hard work to change how the medical centers operate. And the 
result is breaking the EHR as well as the medical centers where 
they actually are functioning.
    What do you think is going to happen if they continue down 
that road?
    Mr. Levin. Mr. Chair, I think that this is a shared 
responsibility. If you look at the implementation frameworks at 
other large clinics--the one that I have in mind specifically 
is Mayo, but there are many of them that we could talk about--
there is a halfway point.
    It does not have to be all one way, my way or the highway--
there are some vendors that do it that way--or the way that we 
have done it at the VA, where it is sort of let a thousand 
flowers bloom, right, until the moment that we are trying to 
modernize the program and exchange data. Even between VistA 
systems, as Secretary Baker described, we can not do that.
    The VA, I believe, is actually primarily responsible. I do 
not want to exonerate any vendor here, right, but the VA is 
responsible, should have been responsible, for making sure that 
they had consolidated those clinical care pathways. I think the 
canonical number is 6, is what we are looking for now, but not 
131.
    Mr. Gfrerer. I think, Mr. Chairman, when you look at the 
capability sets that the EMR vendor or her vendor is going to 
provide, I think there is actually fairly good agreement 
between VHA and the vendor in terms of what compromises those 
capability sets.
    Mr. Rosendale. Thank you so much.
    Representative, do you have any more questions?
    Mrs. Cherfilus-McCormick. No.
    Mr. Rosendale. Then we are all clear. Okay.
    I want to thank our witnesses on both panels for joining us 
today. I do appreciate it. It was incredible information.
    This is not a conversation about IT systems. When we get 
caught up in that, we lose sight of our purpose. It has to be a 
conversation about whether the VA healthcare is meeting our 
veterans' needs and what policies and systems support that.
    It also cannot be a conversation about hypotheticals. It 
has to be grounded in hard data. I appreciate our witnesses 
helping us better understand that data.
    By every measure--by every measure that has been presented 
here today, VistA is still performing well. The VA is still 
very dependent upon it, even after being put on a starvation 
diet in the early years of Cerner. By choice or by default, the 
VA will probably continue to rely on VistA for years into the 
future.
    I realize that not everyone likes that reality, but we all 
need to recognize it. It certainly would not have been my 
choice to operate two different EHRs in a healthcare system, 
but I think it would do a tremendous disservice to the veterans 
who rely on the VA for their care to neglect or even dismantle 
the only platform that supports that care effectively and 
safely.
    We need to be identifying the VistA improvements that are 
most badly needed and make the most sense and get them done in 
months or in years, not in decades. That is what I will be 
advocating.
    Thank you all again for your participation in today's 
hearing.
    I ask unanimous consent that all members have 5 legislative 
days to revise and extend their remarks and exclude extraneous 
material.
    Without objection, so ordered.
    Mr. Rosendale. This hearing is adjourned. Thank you.
    [Whereupon, at 5 p.m., the subcommittee was adjourned.]

     
      
=======================================================================


                         A  P  P  E  N  D  I  X

=======================================================================


                    Prepared Statement of Witnesses

                              ----------                              


                  Prepared Statement of Daniel McCune

INTRODUCTION

    Good Afternoon, Chairman Rosendale, Ranking Member Cherfilus-
McCormick, and distinguished Members of the Subcommittee. Thank you for 
the opportunity to testify today about the Department of Veterans 
Affairs' (VA) Veterans Health Information Systems and Technology 
Architecture (VistA). I am accompanied today by Charles C. Hume, Chief 
Informatics Officer, Veterans Health Administration, Dr. Thomas 
O'Toole, Deputy Assistant Under Secretary for Health for Clinical 
Service, Veterans Health Administration, Ms. Zhuchun ``Emily'' Qiu, 
Director of Health Informatics, Office of Information and Technology, 
and Mr. Michael Giurbino, Director, Health Infrastructure and Systems 
Management, Office of Information and Technology.

OVERVIEW

    VA is committed to providing exceptional care, services, and a 
seamless, unified experience to Veterans. The Office of Information and 
Technology (OIT) collaborates with the Veterans Health Administration 
(VHA) and various VA offices to achieve this mission through the 
delivery of state-of-the-art technology, including a modernized 
Electronic Health Record (EHR).
    Today, VistA and its integrated systems provide an integrated EHR 
for Veteran care and services, supporting over 150 applications, 
including the operations of more than 1,500 VA facilities. There are 
133 instances of VistA nationwide that share standard functionality but 
have data and workflow tailored to the needs of each VA Medical Center 
and its patient population. Like any IT system, VistA requires updates 
and maintenance to keep it functioning at a high level. Critical 
upgrades to the system could be extremely costly over the years, and 
maintenance costs are even higher. Often, it becomes more expensive to 
maintain a legacy system than to replace it. VistA itself is written in 
an old programming language, Mumps. There are few Mumps programmers 
today, Mumps is not taught in computer science classes, and the pool of 
Mumps programmers shrinks every year as they retire. VA is fortunate to 
have dedicated Mumps programmers supporting VistA. They understand 
millions of lines of code developed over 45 years and VA's clinical 
business processes. They are committed to enabling clinicians and 
supporting Veteran outcomes. We've been able to retain them, and their 
knowledge, much longer than a typical workforce. However, approximately 
70 percent of our Mump programmers today are retirement eligible, and 
we have few options to hire or contract additional ones.
    VistA has served VA and Veterans for over 45 years and we aware of 
its limitations. It doesn't have modern capabilities like Artificial 
Intelligence/Machine Learning, mobile and web access, and capabilities 
providers and Veterans expect and deserve from a modern cloud-native 
EHR. VistA is a member of VA's expansive and complex ecosystem of 
software and infrastructure. The size and complexity of that technology 
ecosystem has nearly doubled in the last 5 years, and most of that 
growth was in modern cloud-native applications. Mumps programmers are 
increasingly challenged keeping VistA integrated in a growing ecosystem 
that is architected very different from the system designed 45 years 
ago. While technology is a challenge, so also are the dated skills of 
the VistA programmers. These challenges compound every year.
    To modernize VA's legacy EHR systems and achieve interoperability 
with DoD and community care providers, VA is transitioning to a new EHR 
solution. In May 2018, VA awarded Cerner a contract to replace VistA 
with a Commercial Off the Shelf (COTS) solution, Cerner Millennium, 
which is also currently being deployed by DoD. During implementation of 
the new EHR solution, VA will need to maintain VistA systems for a 
period of time. This ensures that current patient records remain 
accessible and that there will be no interruption in the delivery of 
quality care.

FUTURE VISTA DEVELOPMENT

    VA recognizes the planned Electronic Health Record Modernization 
(EHRM), Financial Management Business Transformation (FMBT), and Supply 
Chain Modernization (SCM) efforts will take years to scale across the 
enterprise. During this time, maintaining Vista is necessary to ensure 
VA preserves the standard of care in the interim and continues 
innovation to serve the Veteran. VA embraces the responsibility to 
consistently and constantly drive modernization and look for efficient 
ways to sustain VistA. Some of the key strategies include:

      Development, Security, and Operations Approach - OIT 
shifted to a DevSecOps approach focused on collaboration, innovation, 
agile principles, and automation--so that it can develop, enhance, 
maintain, and roll out better, more secure products at a faster pace 
than using the existing separate development and operations processes.

      VistA Standardization - VAMCs are required to run the 
nationally released ``Gold'' version of VistA. In addition to having a 
common set of software routines for each VistA instance, there are some 
additional normalization activities, including work on terminology 
extensions, to account for local differences that will need to be 
addressed to ensure standardization of the VistA data base and file 
system.

      Merging Resources - OIT continues merging VistA teams and 
resources for maximum efficiency throughout VA.

      Maintain excellent customer support - OIT will continue 
to respond to patient safety issues; hiring and retention of VistA 
support resources; maintaining security and compliance (scans, 
remediation, Section 508 compliance, Authority to Operate, etc.); 
refreshing hardware (life-cycle upgrades, hardware, cloud migration 
etc.); and maintaining software versions and upgrades.

    VistA enhancements require enabling teams to work in a development 
paradigm using modern tools and practices such as automation of 
testing, integration, and deployment of code. VistA enhancements share 
VistA data and applications through Application Programming Interfaces 
(APIs) that use modern messaging standards. This approach accelerates 
integration and supports innovation in the short term. It also 
facilitates migration to target solutions like EHRM, FMBT and SCM in 
the long-term.

CLOUD MIGRATION

    On June 22, 2019, one instance of VistA, at Valley Coastal Bend, 
was successfully migrated to VA's Enterprise Cloud (VAEC) which is the 
future direction for VistA instance maintenance until they are subsumed 
by Cerner Millennium. Since then, a total of 20 VistA sites were 
successfully migrated to the VAEC and an additional 54 VistA site 
migrations are planned for Fiscal Year 2023. VA is taking advantage of 
cloud-based infrastructure management practices and leveraging cloud 
native features including security, monitoring, backups, and 
scalability. As part of the current VistA Cloud Migration Project, the 
VistA software platform is also being upgraded to IRIS for Health 
2022.1.

COSTS OF SUSTAINMENT

    For the purposes of ensuring uninterrupted health care delivery, VA 
will continue to use VistA until all legacy systems are replaced by the 
new solution. Below are the current costs to operate, maintain, and 
upgrade VistA in each of the last five fiscal years. The below costs 
reflects a steady increase year-over-year:

      Total Fiscal Year 2018: VistA cost $417,730,309

      Total Fiscal Year 2019: VistA cost $634,138,491

      Total Fiscal Year 2020: VistA cost $720,312,589

      Total Fiscal Year 2021: VistA cost $841,426,084

      Total Fiscal Year 2022: VistA cost $890,098,856

    Currently, there is no VistA sustainment cost reduction directly 
tied to the new EHR solution rollout. VistA must run without service 
degradation to support EHR migration and overall VA modernization. 
VistA clinical modules that are deemed redundant when the EHR migration 
is complete will be decommissioned. However, VistA modules that are not 
replaced by the EHR must be performant and maintained until replacement 
capabilities are developed. The cost to maintain VistA will increase as 
we must include development for new capabilities and interfaces, 
congressional mandates, cloud costs, hiring and retention of VistA 
support resources, and maintenance. To continue fulfilling our 
commitment to ensure uninterrupted care and benefit delivery to 
Veterans, VA must continue to use VistA.

CONCLUSION

    As VistA functionality is replaced by a COTS solution and other 
systems, VA can decommission VistA products as appropriate. Until the 
new EHR solution is implemented across VA's enterprise, VistA remains 
VA's authoritative source of Veteran data. Sustaining VistA for the 
duration of our EHRM effort ensures that Veterans continue receiving 
uninterrupted care and services while VA looks to the future and 
improves the Veteran experience.
    Chairman Rosendale, Ranking Member Cherfilus-McCormick, and Members 
of the Subcommittee, thank you for the opportunity to appear before you 
today to discuss OIT's progress toward VistA transition. I look forward 
to continuing working with this Subcommittee and address our greatest 
priorities. This concludes my testimony, and I look forward to 
answering your questions.
                                 ______
                                 

                   Prepared Statement of Roger Baker

    Thank you Chairman Rosendale and Ranking Member Cherfilus-McCormick 
for holding this hearing today. With over $50 billion at stake, 
misinformation regarding VistA has been rampant. Numerous parties, 
largely those with no expertise in VistA, medical care, or software 
development, continue to repeat misinformation designed to convince 
Congress that VistA is a problem in need of their solution.
    Not that I need to remind this committee, but I served 4 years as 
the Assistant Secretary and CIO for VA, from 2009 to 2013. In that 
role, I was responsible for all investments in VistA, including 
budgeting, daily operations, bug fixes, improvements, modernizations, 
and strategic direction. I was responsible for analyzing and then 
stopping the failed HealtheVet program. I was the VA lead for the iEHR 
program and dealt with all of its complexities. Perhaps most useful for 
our discussion today, I analyzed why VA continued to fail in its 
software development efforts, and introduced a program management 
approach that increased the rate of on-time software deliveries across 
the development portfolio from under 30 percent to over 84 percent.
    Several years ago, I published three articles regarding the EHRM 
program and VistA, which I have linked below. I believe the information 
in these articles is still largely relevant.

    Why VA's Electronic Health Record Mega Project is Failing https://
fcw.com/it-modernization/2021/07/why-vas-electronic-health-record-mega-
project-is-failing/259229/

    How VA Can Succeed with its EHR Mega-Program https://fcw.com/
acquisition/2021/08/how-va-can-succeed-with-its-ehr-mega-program/
258949/

    Why VA Must Keep VistA Healthy https://fcw.com/acquisition/2021/08/
why-va-must-keep-vista-healthy/259006/

    The primary success measurement for an electronic health record 
(EHR) system at VA should be Veteran health outcomes. The fundamental 
reason that VA, VHA, and the EHR exist is to provide Veterans with 
superior health care. An EHR should be an aid to clinicians and medical 
staff in doing their jobs, help speed their work, provide them with 
information to make better decisions, reliably communicate work orders 
such as lab tests, prescriptions, treatments, and specialty referrals, 
and coordinate and track the medical activities of the entire medical 
center needed to improve health outcomes.
    The EHRM program has effectively run a controlled experiment over 
the last six years, complete with a hypothesis, control group, and 
metrics. This experiment has provided concrete proof that Veterans 
achieve better medical outcomes when VA facilities use VistA than when 
they use the commercial alternative. And while service impacts should 
be expected in the initial days of an EHR swap, I have seen no 
projections from VA as to when productivity and medical quality 
measurements using the new EHR will exceed those previously seen in the 
same facility using VistA.
    Because Vista excels at medical care, and it sets a high 
comparative bar. Vista, and the work processes encoded in it, was 
designed, implemented, and honed by VA clinicians to do exactly what a 
clinician needs and what a clinician expects. That is both its blessing 
and its curse.

    Independent surveys show that VistA is the most liked EHR by 
clinicians nationally.

    Medscape EHR Report 2014

    Medscape EHR Report 2016: Physicians Rate Top EHRs

    However, as an IT product, VistA is complex and difficult to 
change. But which would we rather have for medical care, a system that 
clinicians love and IT people hate, or one that IT people love and the 
medical staff hates?
    Some lobbyists would have you believe that the IT difficulties of 
VistA are more important than the medical care advantages. That VistA 
is a problem because of its age, complexity, and the language it is 
written in. In fact, at least one lobbyist would have you believe that 
``Vista cannot be made better.''
    This is provably false. During just my 4-year tenure, many 
improvements were introduced to VistA, including bed management, blood 
bank, Pharmacy re-engineering, registries, and numerous others 
including, fairly notably, Cerner Labs.
    In fact, VistA can be difficult to modernize. But the difficulties 
in modernizing VistA stem not from the software itself, but from three 
root causes that come from VA itself.
    First, every time VA has attempted to replace VistA, first from 
2000 to 2009 under HealtheVet, and second from 2017 until now under 
EHRM, VA has prohibited investment in VistA. This has included 
eliminating promising technologies such as VistA Exchange by 
terminating the EHMP program when it was in beta test. The fact that 
you are told that ``VistA cannot be made better'' as an argument for 
why $50 billion needs to be spent to replace it, when the primary issue 
is that VA has cutoff investment in VistA improvements for 16 of the 
last 24 years, is rich indeed.
    Second, years ago VHA made the decision that veterans receive 
better medical care if each VA is allowed to tailor its care to local 
needs. That Veterans in Fort Harrison can have different care needs 
from those in Palm Beach. This local control of medical care is a 
fundamental part of the medical culture of VHA. As you would therefore 
expect, VistA was specifically designed to easily support that local 
customization. I am certainly not qualified to tell you whether 
localization is a good medical decision. For that discussion, I would 
suggest a discussion with Dr. Ken Kizer as well as numerous other 
medical professionals who have given this topic much more thought than 
I ever could. I can tell you that it is where the sound bite ``VistA is 
not a single system, it is 130 separate systems'' comes. Local 
customizations were a celebrated part of VistA development for many 
years, until the advent of the ``Gold Disk'' program. But these local 
customizations are what make designing, programming and testing changes 
to Vista more difficult, because every change must be tested to work 
with each VistA system.
    VA's experience with Cerner Labs is a good example of the effects 
of the VHA culture on VistA. In (approximately) 2007, VA decided to 
replace the VistA Laboratory module with one from Cerner. Integration 
of the cloud version of Cerner Labs into VistA proceeded, and by 2012 
VA had completed a successful introduction of the package at the 
Huntington, WV VA, where it is (to the best of my knowledge) still in 
use today. The functionality was well received, and a plan was 
presented to roll-out across the rest of VA. Unfortunately, that plan 
was wildly expensive and would take many years. When asked why, the 
program team explained that all of the customization parameters in the 
lab package, which had been a significant portion of the development 
work, would need to be re-discovered, re-verified, re-entered, and re-
tested for each VA hospital, since business process and even the names 
used for each drug vary at each VA. As a result, Huntington remains the 
only VA running Cerner Labs (again, to the best of my knowledge).
    As noted above, VA began its ``Gold Disk'' program in 2011, as part 
of the decision to move VistA to an Open Source model. The goal of the 
Gold Disk was to eliminate the variations between VistA instances by 
identifying software differences and working with VHA to agree on which 
business process, and therefore which software modules, could be used 
at all VA's. By 2015, this had reduced variations in the VistA software 
to under 5 percent across all instances. It is my understanding that VA 
has continued this effort and has further reduced software variations, 
possibly to the point of achieving our goal, a single ``Gold Disk'' 
version of VistA to be distributed to all VA facilities.
    Third, Federal pay grades and procurement practices have eroded the 
base of skilled software developers needed to maintain a complex EHR 
product. Capping salaries at GS 13/14 levels for the most skilled 
Federal IT staff has caused them to seek other employment. And VA 
continuously awards contracts for complex VistA improvements to 
companies that lower their prices to win the work, and then cannot 
employ the necessary skills at the rates that were bid. They would 
rather tell VA ``we can't find MUMPS programmers'' than ``we underbid 
the work'' to justify why they failed to deliver.
    VA has repeatedly failed at efforts to replace VistA. HealtheVet, 
iEHR, and now EHRM were each attempts to replace VistA, not to make it 
better. Each failed, in part, because the difficulty in making the 
software better is not in the software, but in the fundamental VHA 
culture. VistA is tightly attuned to that culture, and well liked by 
the medical staff for exactly that reason. Unless and until a decision 
is made that software standardization is more important than local 
control of healthcare, attempts to replace the VistA product with a 
commercial product that does not support that fundamental part of the 
VHA culture are doomed to certain failure.
    Mr. Chairman, there is much misinformation regarding VistA being 
promulgated in an effort to justify the $50 billion needed for the EHRM 
program. I have attempted to address only a few of them. But the EHRM 
program has provided the best proof that they are either wholly or 
partially untrue. After six years, Veterans continue to achieve better 
healthcare outcomes in VA facilities that use VistA versus the 
alternative. That remains the single most important fact you will hear. 
I commend this committee for demanding to deal with the actual facts 
regarding VistA, its role in veteran healthcare, and its ability to be 
modernized, and I look forward to working with you and answering your 
questions as you further search out those facts.
                                 ______
                                 

                  Prepared Statement of James Gfrerer

    Chairman Rosendale, Ranking Member Cherfilus-McCormick, thank you 
for the opportunity today to appear before the Technology Modernization 
Subcommittee with my fellow former VA Technology Panelists to address 
the VA's current Electronic Health Record.
    As a Veteran, I am a patient in the VA health system, and a 
beneficiary in the VA benefits system, and now pre-registered for VA 
burial benefits. And as a more-than 28-year career Marine Infantry 
Officer with 4 combat deployments, I fully empathize with all our 
Veteran men and women who endure both the visible and invisible wounds 
of military service.
    There is much misunderstanding around VA health care in general. VA 
Health care is unlike commercial systems. VA is funded by government 
appropriation versus commercial health systems who operate on a 
business revenue model. In commercial health care, each patient is 
eligible for all services, where in VA eligibility is based-on complex 
service-connected conditions. VA health care is more specialized and 
expansive than commercial systems comprising unique clinical services 
such as prosthetics, long term care, and dental among others. These are 
substantial differences, even as compared to Department of Defense 
Health care, and are the first set of challenges for any commercial EHR 
to be successfully implemented in the Veterans Health Administration 
(VHA).
    The bottom line is that Federal law, regulation, and policy have 
created this unique health system - and the Veteran Health Information 
and Technology System Architecture (VistA) Electronic Health Record is 
representative of those complex and unique business rules. So it may 
come as no surprise that when a commercial EHR programmed for different 
financial frameworks, with significantly different eligibility rules, 
and not addressing unique VA clinical services, that there are problems 
- and problems that can't be overcome by ``change management.'' Without 
substantial customization, no commercial EHR could address the business 
rules that law, regulation, and policy mandate for Veteran Health care. 
So, if you didn't have a business system configured like VistA, you'd 
have to create or heavily customize a system to perform just like it.
    In the remainder of this Hearing, we will get into greater detail 
about VistA, its modernization efforts, and some additional facts and 
misconceptions, but allow me to offer some highlights as a capstone to 
the larger conversation:

      First, VistA is more than an EHR. It is what 
professionals term an Enterprise Resource Planning or ``ERP'' system, 
which has grown over the years to encompass many administrative, 
financial, and other modules. A number of these will live-on, past any 
end of service date for VistA.

      Second, it is not - I repeat - not an ``IT system'' but 
rather a BUSINESS/MISSION system. Why does this matter? First, because 
the ``Business'' - in this case, VHA - must take prime ownership, to 
include the lifecycle management, capital investment, and change 
management, with OIT playing a continue supporting and technical role.

      Third, some would have you believe that VistA has not 
been modernized, but that assertion is predicated on the fallacy that 
modernization can only occur by replacement. Tech modernization as 
defined by Gartner, Forrester, and others, can be achieved in a myriad 
of other ways from rehosting (e.g., moving to the Cloud), refactoring 
(optimizing the existing code), and encapsulating (exposing to APIs) - 
all of which were done to VistA during my VA tenure.

    Also, let me offer that in many respects Veteran Health Care 
business and technology discussions remain mired in 2017. It was in 
this timeframe that the pursuit of a fully longitudinal health record 
was revalidated with the assumption that it must be on the same 
platform in order for this to be achieved. In 2023, with the maturity 
and adoption of Health Information Exchanges and Heath data standards 
such as HL7 and FHIR, that is no longer the case, which raises another 
topline business issue. Which is the greater challenge for VA presently 
- is it DoD/VA interoperability - or is it VA/Community Care 
interoperability?
    In an era of increasing technical debt and mounting technology 
modernization cost, the Congress must determine where the greatest need 
is for precious technology budget. Presently there are roughly 300,000 
active-duty members annually who matriculate from DoD to VA. Last year 
on the Community Care side, VA saw 6 million referrals out of network 
for 36 million episodes of care. To recap - a one-time transfer of 300K 
servicemember records as compared to 6 million referrals with 36 
million appointments - there is no doubt that the latter is the 
substantially larger problem, across thousands of Community Care 
providers, who are on every available EHR, not one single EHR on which 
DoD and/or VA are operating. Community Care is only anticipated to grow 
larger every year, so VA must address it soon.
    Finally, in an era where technology plays and increasingly 
mainstream and critical role in healthcare delivery, VA must begin to 
operate more efficiently and effectively, as do its Commercial and Non-
Profit Health System counterparts, who are well on their way in this 
regard. These systems understand that technology and information 
technology is the success path, and reciprocally, Health Systems can't 
hire there way out of the problem, much as VHA attempts to do every 
year.
    Mr. Chairman, thank you and the Subcommittee for your interest in 
this vital topic, and I look forward to our discussion.
                                 ______
                                 

                   Prepared Statement of Peter Levin

    Thank you for the privilege of testifying before you today 
regarding the Electronic Health Record Modernization effort at the 
Department of Veterans Affairs.
    I am deeply grateful to Chairman Rosendale, Ranking Member 
Cherfilus-McCormick, and members of this subcommittee for the 
opportunity to share with you my perspective on one of the largest 
civilian information technology projects in history.
    Our commitment to our Nation's veterans transcends party lines and 
political ideology. In an era of especially deep ideological divide and 
social tension, I applaud your leadership, Mr. Chairman, in soliciting 
the best ideas and constructive, fact-based perspectives from across 
the spectrum.
    During my time in public service, and under the leadership of 
Assistant Secretary Baker with whom I am delighted to appear this 
afternoon, I had the honor of working on several medical information 
technology systems that are still in use today. Especially relevant to 
this testimony are the Joint Longitudinal Viewer (JLV) (originally 
known as ``Janus'') and the Blue Button personal health record. 
Launched during a Democratic administration, Blue button was warmly 
embraced by the most recent Republican one, too, ``as a fundamental 
component of any effort to empower patients in their healthcare 
decisions.'' And JLV enables hundreds-of-thousands of clinicians to see 
records across platforms every day.
    It is in that context - access to and interoperability of clinical 
data - that I respectfully offer my observations.
    In my opinion, there are three issues before the government 
regarding VistA at Veterans Affairs:

        1) That the billions of dollars already spent on the Cerner 
        implementation will not scale to enterprise-wide clinical care 
        services on the current path, budget, or timeline

        2) That VA can-and-should sustain the data interfaces and 
        connection frameworks already built to send and receive data 
        from MHS GENESIS

        3) And, most important of all, that VA consolidate its current 
        instances of VistA onto a VA-centered clinical workflow, and 
        augment the VistA model to receive data from third-party 
        providers

    Our ability to deploy VHA's nearly 1,300 facilities is hopelessly 
challenged by the incompatibility of those with each other. Leadership 
- then and now - presumed that VistA instances are fundamentally 
congruent, from clinical workflow as well as data interoperability 
perspectives. This is incorrect. VistA instances do not inter-operate, 
and the agency was unsuccessful in catalyzing alignment around a single 
clinical process and record, the sine qua non of a commercial 
deployment.
    Thoughtful members of our community will sensibly ask how it was 
possible to install Cerner at the Military Health Service was 
ostensibly successful, while the VA's has been a failure. There are 
three reasons. First, DoD had already made the transition to a 
centrally administered system, as opposed to VA's decentralized 
approach that perpetuates workflow (and data model) autonomy across 130 
hosts. Second, the DoD's ``command and control'' structure not only 
enforced protocol alignment, every deployment was preceded by careful 
preparation, training, and integration. Third, their system was in 
substantially worse condition than ours, and Cerner is perceived as an 
improvement over AHLTA, although according to KLAS this is not a widely 
held perception, even at MHS.
    Although VA is one the Nation's largest integrated delivery 
platforms, there are several others in the private sector - including, 
for example, HCA (186 hospitals), Ascension (150 hospitals), and Kaiser 
Permanente (39 hospitals) - each with similarly sized patient 
populations. Before VA embarked on a multi-billion dollar health record 
modernization, we should have been clear on the price-and-performance 
benchmarks from near peer enterprises. We should well understand the 
capabilities they prioritize in their information systems before we try 
to install one ourselves. As you'll hear from other witnesses, the 
differences in VA healthcare between cities is simply not that large. 
It is not tens-of-billions-of-dollars large.
    The inescapable solution to the real-world challenges of first-
class healthcare services at VA is an institutional commitment to 
rebalance the ``have it our way'' approach. VA must truly, sincerely, 
authentically, put the Veteran first, and streamline its own processes 
before it attempts to automate them. As straightforward as this sounds 
to anyone who works for (or is a customer of) a manufacturer, a school, 
a hospital, a store, a publisher, or a transportation, energy, or 
services company, it is exactly not what happens at VA. Veterans 
receive terrific healthcare, but their care is delivered with different 
processes depending on which hospital or clinic they go to.
    There is surprisingly little in the operational literature about 
how to drape an enterprise management framework over complex clinical 
environments, never mind one as large and diverse as VA's.
    From a data perspective, this has profound implications. When we 
try to connect applications-to-pipelines-to-governance, the 
transactional perspective and the analytical one are, literally, 
geometrically orthogonal to each other. Transactions are row based 
(every new interaction, like delivering a vaccine, or serving a meal, 
requires a new entry) and analytics are column based (so as to avoid 
the need to ingest every attribute of every transaction for every 
report, like how many injections were made, and how many meals were 
served).
    In practical terms, while VistA communicates between members of the 
care team one patient at a time, all of its data is exported to a 
separate system to measure outcomes and improve service. The current 
effort to replace both components of that at once is difficult to do. 
We're trying to fit round transactional pegs into square analytical 
holes. We should stop doing that.
    In my opinion, the department should not announce its intention to 
change the contract unless and until it has a backup plan in place. 
That plan cannot be ``revert back to VistA'' in its current form, or 
anything that concedes to VA's continued digital isolation and process 
insularity. Yes, Cerner has performance deficits in response time, 
uptime, and data syndication latency; these will only get worse as more 
hospitals are brought online.
    The problems are compounded, however, by VA's inability to prepare 
for Cerner's deployment. VA in general relies too much on industry to 
``tell it what it needs,'' never mind what it could have. Unless and 
until the institution is committed to aligning its internal processes, 
no amount of technology will automate its operations, and no amount of 
money will solve its policy problems.
    The trade-space is simple: VA must go on the record and publicly 
State their commitment to a single enterprise--workflow and--data 
model. Congress could develop objective and quantitative measures to 
validate compliance, with real consequences for noncooperation or 
nonparticipation.
    If we change the agreement with the commercial vendor, VA must also 
prepare itself for an onslaught of criticism because of previous 
attempts that also failed, including HealtheVet, iEHR, and Vista 
Evolution. Switching back to VistA and walking away does not fix the 
root problems. How do we address the issue?
    First and foremost, cloud technologies are now stable and mature 
enough to enable consolidation onto an authentically single platform. 
Additionally, there have been substantial improvements to the codebase 
that are now available to VA from the commercial sector. I believe that 
the Open Source Electronic Health Record Alliance was the right idea, 
but it was poorly implemented because of VA's (and DoD's) lack of 
sustained commitment and the peculiarity of a ``single customer'' 
market. Nonetheless, it would be straightforward to re-instantiate 
OSEHRA with a powerful charter and legislative mandate. It would 
certainly be objectively better, and cheaper, than what we have now.
    Moreover, commercially available data management infrastructure has 
made substantial progress in the last 5 years. The VA has been 
dreadfully slow to adopt these standards and tools, including FHIR and 
bi-directional Blue Button, because of VistA's inability to ingest 
third-party clinical data into the enterprise model. There are no 
technology impediments here; this is simply a matter of political will, 
imaginative leadership, and execution accountability.
    Indeed, both these policy changes - consolidation onto an 
enterprise clinical workflow, and adoption of proven platform and data 
management services - would accelerate health record modernization at a 
fraction of the costs now earmarked for EHRM. They can be 
systematically procured, thoroughly tested, and methodically deployed 
during this Congress and sustainably thereafter.
    Software is designed to help automate repetitive tasks that we do 
every day. If what clinicians do every day is different at every 
hospital, and we allowed those points of care to grow their own for 
decades, it is no surprise that all the software is going to be 
different. How we got here is no mystery.
    If VA shared best practices between hospitals, identified optimal 
workflows, disseminated them to their network, and updated the software 
unto a unified platform, then clinicians throughout the enterprise 
would provide care in a similar (and probably better) way. This would 
not only improve safety and outcomes, it would be more amenable to a 
wholesale replacement.
    The transcendent goal is continuously better healthcare for 
Veterans. Until VA resolves its internal tension around a consolidated 
workflow and comprehensive data management, no new or renovated 
electronic record will be successful.

                       Submission for the Record

                              ----------                              


 Office of Inspector General Letter to The Subcommittee on Technology 
                             Modernization

[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]


                                 [all]